Page 1
University of Central Florida University of Central Florida
STARS STARS
Electronic Theses and Dissertations, 2004-2019
2018
A Study on the Plasticity and Fracture of AISI 4340 Steel Under A Study on the Plasticity and Fracture of AISI 4340 Steel Under
Different Loading Conditions Considering Heat Treatment and Different Loading Conditions Considering Heat Treatment and
Micromechanics Micromechanics
Sami Ghazali University of Central Florida
Part of the Mechanical Engineering Commons
Find similar works at: https://stars.library.ucf.edu/etd
University of Central Florida Libraries http://library.ucf.edu
This Doctoral Dissertation (Open Access) is brought to you for free and open access by STARS. It has been accepted
for inclusion in Electronic Theses and Dissertations, 2004-2019 by an authorized administrator of STARS. For more
information, please contact [email protected] .
STARS Citation STARS Citation Ghazali, Sami, "A Study on the Plasticity and Fracture of AISI 4340 Steel Under Different Loading Conditions Considering Heat Treatment and Micromechanics" (2018). Electronic Theses and Dissertations, 2004-2019. 5881. https://stars.library.ucf.edu/etd/5881
Page 2
A STUDY ON THE PLASTICITY AND FRACTURE OF AISI 4340 STEEL UNDER
DIFFERENT LOADING CONDITIONS CONSIDERING HEAT TREATMENT AND
MICROMECHANICS
by
SAMI MANSOUR GHAZALI
M.S. University of Central Florida, 2013
B.S. California State University, 2011
A dissertation submitted in partial fulfillment of the requirements
for the degree of Doctor of Philosophy
in the Department of Mechanical and Aerospace Engineering
in the College of Engineering and Computer Science
at the University of Central Florida
Orlando, Florida
Spring Term
2018
Major Professor: Yuanli Bai
Page 3
ii
Β©2018 Sami Ghazali
Page 4
iii
ABSTRACT
Accurate predictions of material strength under different loading conditions with large
plastic deformation and ductile fracture are of great importance. This dissertation aims to develop
an essential understanding of ductile fracture of AISI 4340 steel alloy using both empirical and
micromechanics based models. For this purpose, 29 specimens of different geometries with
different heat-treatments were designed to investigate the effects of stress states. These specimens
are: (a) 13 round bars with different notches (axial symmetric tension); (b) 13 flat grooved
specimens with different grooves; (c) 3 small round cylinders. Mechanical tests up to fracture were
conducted on these specimens to characterize the influence of hydrostatic stress and Lode angle
on material plasticity and fracture. Scanning electron microscopy (SEM) observations were
performed on both original and fractured specimens to investigate different micromechanical
revelations and features. The plasticity model with pressure and Lode angle effects (PPL model)
and the modified Mohr-Coulomb (MMC) fracture criterion were used to predict plastic flow and
fracture initiation behaviors under different loading conditions in finite element simulations. A
model optimization method using ISIGHT was set up to achieve simulation results that were well
correlated with experimental data. The effects of heat-treatment on material strength and ductility
of AISI 4340 steel were also discussed. This work was further carried onto the microvoids based
metal plasticity theory. The well-known Gurson-Tvergaard-Needleman (GTN) model was
calibrated for AISI 4340 steel. It is found that the GTN model is not sufficient in simulating test
data for the 16 Rockwell hardness plane strain specimens. Therefore, The GTN model is modified
to include the Lode angle dependence on matrix material plasticity. It is also found that using a
Page 5
iv
fixed or constant microvoid volume fraction at failure (ff) for all loading conditions is inadequate.
Following a similar derivation of the MMC fracture model, the microvoid volume fraction at
failure (ff) becomes a function of both stress triaxiality and Lode angle. This new criterion is named
(GTN-MMC). The proposed plasticity and fracture models were implemented into ABAQUS
through a user-defined material subroutine (VUMAT) for FE simulations. Good correlations were
achieved between experimental results and numerical simulations.
Page 6
v
ACKNOWLEDGMENTS
First, I would like to thank my advisor, Professor Yuanli Bai for his great mentoring and
guidance throughout my graduate training at the University of Central Florida. I am very
appreciative of the daily meetings and constructive feedback that enabled this milestone of the
degree to be reached successfully. Dr. Bai, without your fruitful comments and suggestions, this
would have been difficult.
I also would like to thank my committee members Professor Alain Kassab, Professor Kawai
Kwok and Professor Boohyun Nam for their fruitful comments, discussions, and support. I am
also grateful to all my colleagues who supported me, especially Dr. Yueqian Jia, Dr. Mohammed
Algarni and Dr. Yangyang Qiao for their endless assistance and generous feedback.
I would also like to express my profound thanks and appreciation to Timothy Linder (UCF
machine shop manager) who aided in constructing all tested specimens. I would like to thank the
College of Engineering at the University of Jeddah for granting me a full scholarship to peruse
my Ph.D. education in the United States. I would also like to thank the Saudi Arabian
Government for granting me King Abdullah Bin Abdulazizβs scholarship to peruse my Bachelor
and Master degrees at the United States of America.
My sincere gratefulness goes to my family; especially my parents, sisters, brother, wife, son, and
daughter for their genuine encouragement, patience, and prayers.
Page 7
vi
TABLE OF CONTENTS
LIST OF FIGURES ........................................................................................................................ x
LIST OF TABLES ....................................................................................................................... xiv
CHAPTER 1: INTRODUCTION ................................................................................................... 1
1.1 Background .................................................................................................................... 1
1.2 AISI 4340 Steel Alloy and Experiments ........................................................................ 2
1.3 Heat-treatments .............................................................................................................. 4
1.4 Dissertation Overview .................................................................................................... 6
CHAPTER 2: LITERATURE REVIEW ........................................................................................ 7
2.1 Empirical Based Models ................................................................................................ 7
2.1.1 Johnson-Cook Model (J-C) .................................................................................8
2.1.2 Wilkins Model .....................................................................................................8
2.2 Phenomenological Based Models .................................................................................. 9
2.2.1 Maximum Shear Stress Model ............................................................................9
2.2.2 Pressure Modified Maximum Shear Stress Theory ...........................................10
2.2.3 Cockcroft-Latham Model ..................................................................................10
2.3 Physics-Based Model (Micro-Mechanical Models) ..................................................... 11
2.3.1 McClintock Model ............................................................................................11
2.3.2 Rice-Tracey Model ............................................................................................12
Page 8
vii
2.3.3 Thomasonβs Model ............................................................................................12
2.3.4 Brown and Embury Model ................................................................................13
2.3.5 Bouaziz and Maire Criterion .............................................................................13
2.3.6 Gurson Type Models .........................................................................................14
2.4 Microvoid Behaviors .................................................................................................... 15
2.4.1 Void Nucleation ................................................................................................17
2.4.2 Void Growth ......................................................................................................19
2.4.3 Void Coalescence ..............................................................................................20
2.4.4 Void Closure and Healing .................................................................................20
2.4.5 Void Size and Shape Effects .............................................................................21
CHAPTER 3: EXPERIMENTAL METHODS AND SPECIMEN DESIGN ............................. 23
3.1 Material Specification and Heat-treatment ................................................................... 23
3.2 Representation of Stress State ...................................................................................... 25
3.3 Design of Specimen Geometries .................................................................................. 27
3.4 Theoretical Analysis of Stress States ........................................................................... 32
3.5 Experimental Setup ...................................................................................................... 35
CHAPTER 4: PLASTICITY AND FRACTURE MODELS CALIBRATION AND NUMERICAL
VERIFICATION.......................................................................................................................... 38
4.1 The Drucker and Prager Plasticity Model .................................................................... 38
4.2 The PPL Model and Failure Criterion .......................................................................... 39
Page 9
viii
4.2.1 Modification to the Original Bai and Wierzbicki Plasticity Model ..................39
4.2.2 The Deviatoric Associated Flow Rule of the Modified PPL Model .................42
4.2.3 Modified Mohr-Coulomb Criterion Failure Criteria .........................................43
4.3 Calibration Procedure ................................................................................................... 44
4.3.1 Calibration of the Drucker and Prager Plasticity Model ...................................46
4.3.2 Calibration of the PPL Plasticity Model ............................................................47
4.4 Comparison of Experimental and Numerical Simulations Results .............................. 49
4.4.1 Plastic Strength and Flow of AISI 4340 Steel with Different Hardness ...........51
4.4.2 Fracture Surface Determination ........................................................................65
4.5 Results Discussion of the PPL Model Coupled With MMC Fracture Criteria ............ 70
CHAPTER 5: EXTENSION AND APPLICATION OF MICROMECHANICS BASED
GURSON TYPE MODEL ............................................................................................................ 72
5.1 Model Framework and Previous Studies ...................................................................... 72
5.2 Theoretical Improvements to the GTN Model ............................................................. 82
5.2.1 Modified GTN Plasticity Model with Lode Angle Matrix Material
Dependence ...................................................................................................................83
5.2.2 Modified Evolution Rule of Microvoid Volume Void Fraction (π) .................84
5.2.3 Modified Criterion of Critical Microvoid Volume Fraction to Fracture (ππ)
Depending on Different Stress States ............................................................................85
Page 10
ix
5.3 Experimental Results of AISI 4340 Steel and Numerical Validation of the Proposed
(GTN-MMC) Model ..................................................................................................... 89
5.3.1 Validation of Proposed Model (GTN-MMC) ...................................................96
5.3.2 Plasticity Model Calibration ..............................................................................96
5.3.3 Calibration of Failure Criteria (GTN-MMC) ....................................................97
5.3.4 Comparison between Tests and Simulation Results ..........................................98
CHAPTER 6: STUDY OF DUCTILE FRACTURE FROM MATERIAL
MICROSCOPIC IMAGES ....................................................................................................... 102
CHAPTER 7: CONCLUSION AND FUTURE WORK ....................................................... 121
7.1 Summary of Contributions ......................................................................................... 121
7.2 Recommended Future Study ...................................................................................... 124
: FLOW RULE DERIVATION CONSIDERING ππ(π½) ........................... 127
: DETERMINING TRUE STRESS-STRAIN CURVES BASED ON
BRIDGMANβS ROUND NOTCHED BARS USING DIGITAL
IMAGING CORRELATION (DIC) ........................................................................................... 134
: PLASTIC FLOW POTENTIAL DERIVATION FOR PROPOSED
MODIFIED GTN MODEL......................................................................................................... 148
: HEAT-TREATMENT CERTIFICATIONS .............................................. 151
REFERENCES ........................................................................................................................... 154
Page 11
x
LIST OF FIGURES
Figure 2-1: Evolution of the damage process in a ductile material that is under tensile loading
showing different stages of void behavior (Abbassi, PantalΓ© et al. 2010) .................................... 16
Figure 2-2: Damage in aluminum 6061matrix reinforce with Al2O3 (a) decohesion; (b) particle
fracture (Kanetake, Nomura et al. 1995) ...................................................................................... 17
Figure 2-3: Void closure steps forming a crack (Zhang, Cui et al. 2009) .................................... 21
Figure 3-1: Principal stresses in spherical coordinate representation (Bai and Wierzbicki 2008)
....................................................................................................................................................... 26
Figure 3-2: Round bars and flat grooved specimens in the unheated case (16 HRC) .................. 30
Figure 3-3: samples of other round bar specimens with different heat-treatment conditions ...... 30
Figure 3-4: Drawings of the small compressive cylinder ............................................................. 31
Figure 3-5: Drawings of round bars (SB, LN, and SN) ................................................................ 31
Figure 3-6: Drawings of flat grooved specimens (FLG, FMG, and FSG) .................................... 32
Figure 3-7: Test configuration setup ............................................................................................. 36
Figure 3-8: Samples of round bar fractured specimens (cup and cone fractured surface) ........... 37
Figure 3-9: Samples of flat grooved fractured specimens (slant fractured surface) ..................... 37
Figure 4-1: Comparison of stress-strain curves of heated (32 Rockwell) and unheated (16
Rockwell) specimens .................................................................................................................... 46
Figure 4-2: Samples of meshes used in Abaqus for simulations .................................................. 50
Figure 4-3:Comparison of plastic deformations between test and simulations for un-heated (16
HRC) round specimens ................................................................................................................. 51
Page 12
xi
Figure 4-4:Comparison of plastic deformations between test and simulations for heat treated (32
HRC) round specimens ................................................................................................................. 52
Figure 4-5: Comparison of plastic deformations between test and simulations for heat treated (39
HRC) round specimens ................................................................................................................. 53
Figure 4-6: Comparison of plastic deformations between test and simulations for un-heated (16
HRC) flat grooved specimens ....................................................................................................... 54
Figure 4-7: Comparison of plastic deformations between test and simulations for heat treated (32
HRC) flat grooved specimens ....................................................................................................... 54
Figure 4-8:Comparison of plastic deformations between test and simulations for heat treated (39
HRC) flat grooved specimens ....................................................................................................... 54
Figure 4-9: Comparison of plastic deformations between test and simulations for un-heated (16
Rockwell) small compressive cylinder ........................................................................................ 55
Figure 4-10: Force-displacement curves of AISI 4340 steel (16 HRC) using DP model ............. 58
Figure 4-11: Force-displacement curves of AISI 4340 steel (32 HRC) using DP model ............. 59
Figure 4-12: Force-displacement curves of AISI 4340 steel (39 HRC) using DP model ............. 60
Figure 4-13: Force-displacement curves of AISI 4340 steel (16 HRC) using the PPL model ..... 61
Figure 4-14: Force-displacement curves of AISI 4340 steel (32 HRC) using the PPL model ..... 62
Figure 4-15: Force-displacement curves of AISI 4340 steel (39 HRC) using PPL model ........... 63
Figure 4-16: Comparison of true stress-strain curves from experimental and FEA results for
untreated AISI 4340 steel small compressive cylinder (16 HRC) ............................................... 64
Page 13
xii
Figure 4-17: Calibrated 2D fracture locus of axial symmetric tension (round bar specimens) at =
1 and plane strain tension (flat grooved specimens) at = 0 ...................................................... 66
Figure 4-18: Calibrated 3D fracture locus of unheated AISI 4340 steel (16 HRC). .................... 67
Figure 4-19: Calibrated 3D fracture locus of heat-treated AISI 4340 steel (32 HRC). ................ 68
Figure 4-20: Calibrated 3D fracture locus of heat-treated AISI 4340 steel (39 HRC). ................ 69
Figure 4-21: Shapes of the initial yield surfaces at the Ο plane for AISI 4340 steel with different
hardness......................................................................................................................................... 71
Figure 5-1: Experimental load-displacement curve and the corresponding FEA results with J2
plasticity model (Kiran and Khandelwal 2014). ........................................................................... 81
Figure 5-2: Hardening curves for 16 HRC, 32 HRC and 39 HRC consecutively used in GTN model
....................................................................................................................................................... 89
Figure 5-3: Force-displacement curves for (16 HRC) GTN model .............................................. 92
Figure 5-4: Stress-strain curve of the small compressive cylinder (16 HRC) .............................. 93
Figure 5-5: Force-displacement curves for (32 HRC) .................................................................. 94
Figure 5-6: Force-displacement curves for (39 HRC) .................................................................. 95
Figure 5-7: Calibrated 2D fracture locus of axial symmetric tension (round bar specimens) at =
1 and plane strain tension (flat grooved specimens) at = 0 ...................................................... 99
Figure 5-8: Calibrated 3D fracture locus of AISI 4340 steel (16 HRC). .................................... 100
Figure 5-9: Calibrated 3D fracture locus of AISI 4340 steel (32 HRC). .................................... 100
Figure 5-10: Calibrated 3D fracture locus of AISI 4340 steel (39 HRC). .................................. 101
Page 14
xiii
Figure 6-1: Fracture surfaces of round bar specimens for 16 HRC, 32 HRC, and 39 HRC
consecutively............................................................................................................................... 103
Figure 6-2: Fracture surfaces of flat grooved specimens for 16 HRC, 32 HRC, and 39 HRC
consecutively............................................................................................................................... 104
Figure 6-3: Fractured smooth bar specimen at different magnifications (16 HRC) ................... 105
Figure 6-4: Fractured smooth bar specimen at different magnifications (32 HRC) ................... 106
Figure 6-5: Fractured smooth bar specimen at different magnifications (39 HRC) ................... 107
Figure 6-6: Fractured Large Notch specimen at different magnifications (16 HRC) ................. 108
Figure 6-7: Fractured Large Notch specimen at different magnifications (32 HRC) ................. 109
Figure 6-8: Fractured sharp notch specimen at different magnifications (16 HRC) .................. 110
Figure 6-9: Fractured sharp notch specimen at different magnifications (32 HRC) .................. 111
Figure 6-10: Fractured sharp notch specimen at different magnifications (39 HRC) ................ 112
Figure 6-11:Fractured large flat grooved specimen at different magnifications (16 HRC) ....... 113
Figure 6-12: Fractured large flat grooved specimen at different magnifications (32 HRC) ...... 114
Figure 6-13: Fractured large flat grooved specimen at different magnifications (39 HRC) ...... 115
Figure 6-14: Fractured medium flat grooved specimen at different magnifications (16 HRC) . 116
Figure 6-15: Fractured medium flat grooved specimen at different magnifications (32 HRC) . 117
Figure 6-16: Fractured small flat grooved specimen at different magnifications (16 HRC) ...... 118
Figure 6-17: Fractured small flat grooved specimen at different magnifications (32 HRC) ...... 119
Figure 6-18: Fractured small flat grooved specimen at different magnifications (39 HRC) ...... 120
Figure B-1: The cross-section of a notched specimen ................................................................ 135
Page 15
xiv
LIST OF TABLES
Table 3-1: Chemical composition in %wt. of Steel 4340 being tested ......................................... 23
Table 3-2: Round bar specimenβs names and dimensions (mm) .................................................. 28
Table 3-3: Flat grooved specimenβs names and dimensions (mm)............................................... 29
Table 4-1: Achieving different yield criterions by adjusting the PPL parameters (Algarni 2017)
....................................................................................................................................................... 41
Table 4-2: List of material fundamental plasticity properties used in the FEA ............................ 45
Table 4-3: List of material parameters used in the DP plasticity model ....................................... 47
Table 4-4: List of material parameters used in the PPL plasticity model. .................................... 48
Table 4-5: List of material fracture parameters using MMC model ............................................. 48
Table 4-6: Dimensions of round bar specimens (axially symmetric) before and after the test. ... 56
Table 4-7: Dimensions of flat grooved (plane strain) before and after the test. ........................... 56
Table 5-1: Effects of varying GTN model parameters ................................................................. 79
Table 5-2: GTN parameters used for 16 HRC, 32HRC and 39 HRC consecutively .................... 90
Table 5-3: Modified GTN plasticity parameters for 16 HRC, 32 HRC, and 39 HRC consecutively
....................................................................................................................................................... 97
Table 5-4: GTN-MMC fracture criteria ........................................................................................ 98
Page 16
1
CHAPTER 1: INTRODUCTION
1.1 Background
Many engineering designs have inherent flaws that could affect the reliability and
performance intended. Therefore, insufficient safety measurements and poor understanding of
fracture mechanics could result in catastrophic accidents. The use of βfactors of safetyβ is a
standard method in engineering design to account for inherent flaws and defects of materials.
However, this does not give any physical understanding of the problem and may still lead to failure.
Enormous disasters of the twentieth century such as the Boston molasses failure, Comet disaster,
and the Aloha airline Boeing fuselage failure has caused the engineering community to sit and
reevaluate the conventional design approaches. The analysis of these failures emphasized the role
of a solid understanding and implementation of fracture mechanics into design stages (especially
when considering the effects of inherent flaws).
Predictions of different fracture phenomena are essential to validate the proposed plasticity
and fracture models. Many efforts have been taken over the past few decades to model the correct
plasticity and fracture behaviors. The importance of this subject arises in many fields were specific
safety measures should be taken. For example, car manufacturers are always looking for optimal
designs and solutions that are safe, reliable and save money. These designs should always address
the fracture and failure phenomena. Various simulation packages are gaining much popularity due
to recent evolutions in computing power. Successful simulations of any system under extreme
loading events require high fidelity models. Ductile fracture in steels is believed to be a multistep
Page 17
2
process that occurs due to void microscopic activity ranging from growth, nucleation, shear, and
coalescence of microvoids (Garrison and Moody 1987). A solid physical understanding of
different ductile fracture models and criteria are essential to enable the advancement of simulations
in this area. The present dissertation is concerned with empirical and micromechanical based
aspects of ductile fracture of AISI 4340 steel.
1.2 AISI 4340 Steel Alloy and Experiments
A clear understanding of mechanical properties and behaviors of materials are crucial in
predicting failure. Due to the diversity of different materials, finite element method techniques
(FEM) are gaining much publicity in many engineering applications. The use of (FEM) tools and
software allows unnecessary test repetitiveness and thus saving time and money. To be able to use
such tools, a comprehensive elasticity, plasticity, and fracture models become essential. Many
models have been proposed and extensively investigated by the mechanicβs community over the
past four decades; several macroscopic and microscopic models have been developed to describe
the ductile fracture behaviors of metals. These models can be categorized into three groups:
physical based models, phenomenological and empirical models. (Cockcroft and Latham 1968,
McClintock 1968, Rice and Tracey 1969, Gurson 1975, Gurson 1977, Wilkins, Streit et al. 1980,
Johnson and Cook 1985, Wierzbicki, Bao et al. 2005, Nahshon and Hutchinson 2008, Xue 2008,
LauNielsen and Tvergaard 2009, Dunand and Mohr 2011, Bai and Wierzbicki 2015). This
dissertation aims to investigate the applicability of the plasticity, pressure and Lode angle (PPL)
model (An empirically based model developed by (Bai 2008) ) coupled with the modified Mohr-
Coulomb (MMC) criterion in predicting fracture for AISI 4340 steel considering heat-treatment
Page 18
3
effects. Nevertheless, revisit and evaluate the applicability of the Gurson-Tvergaard-Needleman
(GTN) model (physics-based model) in predicting the correct mechanical behaviors under
different loading conditions. This reevaluation of the (GTN) model has resulted in some suggested
modifications to include the Lode angle effects which are not considered in the original derivation.
Nevertheless, a new failure criterion known as (GTN-MMC) is developed.
AISI 4340 steel is a low alloy steel that contains nickel, chromium, and molybdenum. It is
very famous for its high toughness and capability of developing high strength after proper heat-
treatment while still retaining good fatigue strength (Sirin, Sirin et al. 2013). This alloy has many
applications such as commercial and military aircraft landing gears, power transmissions gears and
shafts; nevertheless, it is also used in designing many war projectiles (ASTM-STEEL , HΓ
KAN
2011). The melting point of this tough material is roughly about 1427ΛC which makes good use
in high-temperature environments (ASTM-STEEL). It is also worth noting that this material has
good ductility, especially in the unheated phase.
The classical J2 plasticity theory assumes that hydrostatic pressure and the third deviatoric
stress invariant have no effects on the material yielding and plastic flow. However, many recent
experiments show that these two parameters should be taken into consideration especially when
modeling metals to capture the correct size and shape of a yield surface. (McClintock 1968, Rice
and Tracey 1969, Hancock and Mackenzie 1976, Cazacu and Barlat 2004, Gao, Wang et al. 2005,
Fossum and Brannon 2006, Gao and Kim 2006, Nahshon and Hutchinson 2008, Xue 2008, Gao
2009, Ghajar, Mirone et al. 2012, Gruben, Hopperstad et al. 2012, Liu, Sun et al. 2013, Cao, Gachet
et al. 2014, Gachet, Delattre et al. 2014). In general, stress triaxiality (normalized hydrostatic
Page 19
4
pressure) is a parameter to control the size of the yield surface. On the other hand, the Lode angle
usually controls the shape of the yield surface. The recent empirical model proposed by (Bai and
Wierzbicki 2008) (PPL) considers these two parameters (stress triaxiality, Lode angle parameter)
for both plasticity and ductile fracture. This model is used along with the modified Mohr-Coulomb
fracture criterion under the assumption of proportional loading to predict different fracture
initiation modes. Over the past few years, the modified Mohr-Coulomb has gained much publicity
due to its ability to capture the correct fracture mode (Bai and Wierzbicki 2009, Jia, Ghazali et al.
2017).
1.3 Heat-treatments
Heat treated AISI 4340 steel has been investigated and studied extensively by many different
researchers. (Tanaka and Spretnak 1973, Bell and Loh 1982, Hickey and Anctil 1985, O'Brien and
D 1991, Lee and Su 1999, Odeshi and Bassim 2009, Foster, Chen et al. 2011, Podder, Mondal et
al. 2012, Pang, Li et al. 2013, Sirin, Sirin et al. 2013) . A recent study on this subject was performed
by Sirin and his team (Sirin, Sirin et al. 2013). They investigated the influence of ion nitride on
AISI 4340 steel mechanical and fatigue properties by using both unheated and heat-treated
specimens by austenitizing. The ion nitriding process is done by applying some DC volts between
a container wall that is positively charged and an insulated center post supporting the work to be
nitrided that is negatively charged. Once the job starts a glow of charge of the iodized gas
accelerates and diffuses towards the specimenβs core. The study concluded that initial
microstructure would influence the material response to nitriding. This observation was evident
when comparing the heat-treated specimens with those that are not. The obtained case depths
Page 20
5
deposited in the heat-treated samples were a little more substantial and increased as time and
temperature increased.
Podder (Podder, Mondal et al. 2012) investigated the effects of heat-treatment on flow
formability and mechanical properties of AISI 4340 steel. A linkage between mechanical
properties and the microstructure type was established. It was noted that spheroidized, annealed
and tempered then annealed respectively exhibited an increase in strength for both preforms and
flow formable tubes. On the other hand, a reduction in ductility is observed for all preforms. These
observations confirm the existence of a direct linkage between microstructure and mechanical
properties. The martensite microstructure, typically the hardest steel type microstructure, exhibited
the highest strength and lowest ductility. In contrary, the spheroidized microstructure, typically the
softest of steel microstructure, clearly demonstrates the lowest strength and highest ductility.
To understand the microstructural features of AISI 4340 steel under quenched and tempered
conditions, Lee and Su (Lee and Su 1999) studied the effects of seven different tempering
temperatures as well as quenching using a period of 2 and 48 hours respectively. They observed
that while holding the tempering time constant, the higher the tempering temperature, the less
strength and hardness the material reveals. The only exception to this observation was at a
tempering temperature of 300 ΛC and tempering period of 2 hours. An increase in hardness was
observed. The loss of strength is assumed to be due to the formation of interlath carbide films and
the retained interlath austenite.
The effects of high strain rate on fracture for steel 4340 under different tempering conditions
were investigated by Odeshi and Bassim (Odeshi and Bassim 2009). It was observed and noted
Page 21
6
that thermal softening as a result of adiabatic heating controls the deformation, fracture behavior
and void growth and nucleation. It is worth noting that the dimples observed under the microscope
are narrower for specimens tempered at lower temperatures. Thus, higher tempering temperatures
typically enhance void growth rather than dislocation activities.
1.4 Dissertation Overview
This work is partitioned into seven chapters. CHAPTER 2: is a brief overview of different
fracture models and their implementations. CHAPTER 3: will discuss the experimental methods
and design of the specimens used. CHAPTER 4: illustrates the framework modeling of the Bai-
Wierzbicki, Linear Drucker-Prager models and discussion of their results. CHAPTER 5: discusses
the potential improvements that can be performed on the GTN model. CHAPTER 6: illustrates
some SEM images of fractured specimens at different magnifications. CHAPTER 7: concludes the
research and suggests various topics for future work.
Page 22
7
CHAPTER 2: LITERATURE REVIEW
Several coupled and uncoupled fracture models exist in the literature. (Tresca 1864,
Cockcroft and Latham 1968, McClintock 1968, Rice JR 1969, Gurson 1977, Wilkins, Streit et al.
1980, Johnson and Cook 1985, Thomason 1985, BΓΈrvik, Hopperstad et al. 2001, Bai 2008, Xue
and Wierzbicki 2008, Erice and GΓ‘lvez 2014). The uncoupled models do not incorporate any
damage accumulation parameters in their formulations, while the coupled models do. The
uncoupled models are usually accommodated with independent failure criteria to indicate fracture
initiation and propagation. Many uncoupled models are gaining much publicity due to their ease
of calibration and implementations into various finite element packages. This chapter aims to give
a general overview of different ductile plasticity and fracture theories and models. The models that
are discussed are categorized into three groups: physics-based, phenomenological and empirical
models. Differences between these models are explained and detailed. Finally, different theories
of void behaviors are listed and discussed.
2.1 Empirical Based Models
Empirical modeling refers to models that are created by observation and experiments. These
models do not necessarily have any physical sound. However, they have found many application
implementations due to their ease of calibration and their accurate predictions. This section will
discuss Johnson-Cook and Wilkins models.
Page 23
8
2.1.1 Johnson-Cook Model (J-C)
This model is purely empirical which has many advantages over other models. The J-C
model incorporates the effects of stress triaxiality (normalized hydrostatic pressure), strain rate
and could also include temperature dependency into its formulation (Johnson and Cook 1985). The
J-C model is read as follows:
οΏ½οΏ½ = [π΄ + π΅(νππ)π] [1 + πΆ ln ( νππ
ν0)] (1 β ππ)
(2-1)
where A, B, n, C, ν0 and m are material parameters, and (νππ , ππ) are equivalent plastic strain and
non-dimensional temperature respectively.
ππ =π β ππ
πππππ‘ β ππ
(2-2)
where π, ππ , πππππ‘ are the current, reference and melting temperatures respectively.
Neglecting temperature effects, the J-C fracture model (Johnson and Cook 1985) can also
be defined in the strain domain as follows:
νοΏ½οΏ½ = π0 + π1 exp (βπ2ππΟ) (2-3)
Where π0, π1 and π2 are parameters that need calibration, and ππ, Ο are the mean and equivalent
stresses respectively.
2.1.2 Wilkins Model
This model was developed by (Wilkins, Streit et al. 1980) as a cumulative strain-damage
criterion to predict fracture initiation and propagation. Fracture is activated once damage reaches
a critical value π·πΆ over a critical distance.
Page 24
9
π·πΆ = β«1
(1 β πππ)π
π
0
(2 β π΄)π dΞ΅pl (2-4)
π΄ = max (π 2π 3 ,π 2π 1)
(2-5)
where the A parameter is within 0 β€ π΄ β€ 1 and (π 1, π 2, π 3) are three deviatoric principal stresses.
π is the coefficient for mean stress. π and π are constants that need calibration.
Assuming proportional loading and integrating Eq. (2-4), the fracture locus is expressed as:
νοΏ½οΏ½ = π·πΆ (1 β πππ)
π
(2 β π΄)π
(2-6)
2.2 Phenomenological Based Models
Phenomenological based models are models that define the empirical relationship between
different phenomena. These models have some physical association embedded, but the main thing
is to explain why and how specific variables interact and relate to one another. This section will
discuss maximum shear stress, pressure modified maximum shear stress, and Cockcroft-Latham
models.
2.2.1 Maximum Shear Stress Model
Maximum shear stress theory assumes failure whenever the maximum shear stress in the
designed element reaches a critical value ππ which equals the maximum shear stress in uniaxial
tension. The (Tresca 1864) criterion is one of the most famous criteria used in many civil
engineering applications. It asserts that ππππ₯ = ππ¦ , where ππππ₯ is defined as follows:
Page 25
10
ππππ₯ =ππππ₯β ππππ
2 (2-7)
where ππππ₯ and ππππ are the maximum and minimum principal stresses respectively.
The maximum shear stress is then compared to the yield failure as follows:
ππππ₯β ππππ2
β€ ππ¦ππ (2-8)
The factor of safety (FOS) is defined as:
FOS =ππ¦ππππππ₯
(2-9)
2.2.2 Pressure Modified Maximum Shear Stress Theory
The traditional maximum shear stress theory has no hydrostatic pressure dependence on the
ductile fracture. However, many materials illustrate such reliance. The pressure modified
maximum shear stress reads as follows:
(ππππ₯ + π1 ππ)π = π2 (2-10)
where ππππ₯ =ππππ₯β ππππ
2 and (π1, π2) are material constants. ππ is the mean stress or negative
pressure (βπ).
2.2.3 Cockcroft-Latham Model
The fracture criterion developed by (Cockcroft and Latham 1968) is an increment integral
damage parameter value. This model is most suitable in the range of small and negative hydrostatic
stresses. This criterion has only one material constant that needs calibration which makes it
relatively preferable when sufficient data is absent. The criterion reads as follows:
Page 26
11
π·πΆ = β« < ππππ₯ >π
0
dΞ΅pl (2-11)
where ππππ₯ is the maximum principal stress and νοΏ½οΏ½ is the equivalent strain at fracture. π·πΆ is the
damage material constant to indicate the limitation of the ductile model.
2.3 Physics-Based Model (Micro-Mechanical Models)
Unlike empirical and phenomenological models, physics-based models explain the
microscopic and macroscopic based phenomena in mathematical terms to resemble the ongoing
physical observations.
2.3.1 McClintock Model
(McClintock 1968) analyzed the expansion (growth) and linkage (coalescence) of
cylindrically shaped voids. His main finding is that stress triaxiality is the primary contributor
parameter in governing the fracture behavior. This model focused on the growth of a single void
in an infinite elastic-plastic matrix to determine the critical stain for coalescence. The model reads
as follows:
dD
dΞ΅=
1
ln πΉπ§ππ [
β3
2(1 β π) sinh(
β3 (1 β π)
2 ππ + ππΟ
) +3
4
ππ β ππΟ
] (2-12)
Where ππ , ππ and πΉπ§ππ
are the two principal stresses and the relative void growth factor to
fracture. D is the damage accumulation indicator.
Page 27
12
2.3.2 Rice-Tracey Model
(Rice and Tracey 1969) studied the influence of stress triaxiality on the growth of an isolated
spherical void within an infinite plastic medium. Their model read as:
νοΏ½οΏ½ = π1 exp (βπ2 ππΟ) (2-13)
where π1, π2 are constants that calibration to best describes the fracture locus, and ππ, Ο are the
mean and equivalent stresses respectively.
This model is at most taken as a fracture criterion since it does not consider the effects of
void growth on the material. The Rice and Tracey model has been used in some engineering
applications and was found to give an adequate description of the ductile fracture process (Sun,
Siegele et al. 1989, Zhang 1994).
2.3.3 Thomasonβs Model
Thomasonβs criterion is criterion that assumes a rigid non-hardening plastic solid containing
a regular distribution of spherical microvoids. Unlike many porous models, this model considers
the effects of stress triaxiality on void behaviors. This is done by assuming that the material fails
once his criterion reaches the upper-bound theorem. The void growth rates and geometry changes
are calculated using (Rice and Tracey 1969) void growth model. This model reads as follows
(Thomason 1985):
(πππ)lim
=0.4
(ππ+ππ)2 +
1.67
[(π
π + π) + (
ππ + π
)]
12β
(2-14)
Page 28
13
Where a, b, c, d, and e are geometry dimensions of the void unit-cell system. (ππ
π)lim
is the
incipient limit-load failure of the inter-void matrix based on the current geometry.
(πππ)crit [1 β (
3βπ ππ4
)
23β
(π
π0) (
π
π0) πππΌ] =
πππ+
3 + π£
3βπ£2 + 3
(2-15)
(ππ
π)crit
is the critical condition for incipient ductile fracture based on the upper-bound theorem.
Failure is assumed once the following is satisfied:
(πππ)lim
= (πππ)crit
(2-16)
2.3.4 Brown and Embury Model
(Brown and Embury 1973) proposed a coalescence criterion that is based on internal necking
observed between two adjacent voids. It is assumed that void coalescence initiates once the radius
of the elongated voids is equal to some critical distance between two neighboring void centers.
2.3.5 Bouaziz and Maire Criterion
The (Bouaziz, Maire et al. 2008) criterion is developed with the aid of X-ray tomography
observations. It is assumed that the void density (N) is a function of hydrostatic stress (Ξ·) and
equivalent plastic strain as follows:
π = π΄ (νπππ£π
νπ) exp (
νπππ£π
νπ)
(2-17)
νπ = νπ0 exp (βΞ·) (2-18)
Where π΄ and νπ0 are parameters that onsets the nucleation criterion that needs calibration.
Page 29
14
2.3.6 Gurson Type Models
It is often assumed that hydrostatic stress (stress triaxiality) has little or no effect on yield.
However, studies have shown that the existence of microvoids and impurities will cause
macroscopic dilatation and thus hydrostatic stress should be considered. (Gurson 1977) developed
a yield criterion that links the preexisting of void behavior along with matrix material. The adapted
model is based on a unit-cell of a single void, where the ratio of this void to the cell is defined as π.
Unlike the Rice-Tracy model, Gursonβs model considers not only void expansion and change of
shape, but also the degradation of the materials carrying capacity. However, it was still argued by
many that the model is insufficient to model many realistic materials. Gurson assumed that the
material fails once its porosity reaches unity. This assumption is too big for most materials. (Brown
and Embury 1973) found that once nucleation cavities are formed along the tensile loading, the
two neighboring voids start to coalesce if their lengths are comparable to their spacing (i.e.,
coalescences happens faster than what is assumed by Gurson). Therefore, (Chu and Needleman
1980, Tvergaard 1981, Tvergaard 1982, LauNielsen and Tvergaard 2009) modified Gursonβs work
and included some constant parameters that control the plastic flow potential. They also added a
new parameter known as critical volume fraction to control the coalescence effects. Nevertheless,
a new void nucleation mechanism was developed in addition to void growth. This model is often
referred to as GTN model and is further explained in CHAPTER 5:
Gurson assumed that voids remain spherical regardless of the stress state. However, this is
not the case for low hydrostatic stress, where voids tend to elongate which could have a significant
effect on the prediction of ductile fracture (Benzerga, Leblond et al. 2016). These kinds of
Page 30
15
extensions to the Gurson model were first carried out by (Gologanu, Leblond et al. 1993, CastaΓ±eda
and Zaidman 1994). The reader is highly referred to the recent work of (Benzerga, Leblond et al.
2016) for a comprehensive review of the different Gurson type models and many suggested
improvements.
2.4 Microvoid Behaviors
It has been identified that under high positive stress triaxiality voids grow, nucleate, and
coalesce as illustrated in Figure 2-1. This section aims to shed some light on the former
mechanisms as well as a brief discussion of initial assumptions of void shapes and their effects.
Page 31
16
Figure 2-1: Evolution of the damage process in a ductile material that is under tensile loading
showing different stages of void behavior (Abbassi, PantalΓ© et al. 2010)
Page 32
17
2.4.1 Void Nucleation
X-ray tomography reviles that in heterogeneous nucleation, the most encountered kind,
voids nucleate either by decohesion of the matrix-particle or by fracture of the particle as seen
Figure 2-2 (Puttick 1959, Goods and Brown 1979). Tvergaard argues that nucleation appears to be
controlled by either plastic strain or sometimes by the maximum normal stress when the particle-
matrix reaches a critical value.
Figure 2-2: Damage in aluminum 6061matrix reinforce with Al2O3 (a) decohesion; (b) particle
fracture (Kanetake, Nomura et al. 1995)
By using radiographs on predesigned materials that contain controlled voids drilled by laser,
(Weck 2007) proved that the void nucleation and coalescence strains could be determined. He also
established that the volume void fraction of particles has an inverse relation to both nucleation and
coalescence strains respectively. However, void nucleation is still a challenging subject to model
as it involves many assumptions about the shape and nature of the nucleation, especially when
dealing with real materials. (Tvergaard 1990) argued that these simple shape assumptions are a
reasonable approximation at early stages. However, near-final failure stages theses shape
Page 33
18
approximations are very poor. Landron has illustrated that when using higher X-ray resolutions
voids that appear to nucleate are nothing more than small voids expanding. (Landron 2011)
uncovered that the initial microstructure of the material will affect void nucleation. She observed
that dual phased steels experienced higher nucleation rates when compared to single phase steels.
Void nucleation criterions that are frequently used are the (Argon and Im 1975) ,(Beremin
1981) and (Chu and Needleman 1980) criterions. The Argon criterion is a phenomenological
critical stress criterion that is defined as:
πππ + ππ» = ππΆ (2-19)
where πππ, ππ» and ππΆ are the equivalent, hydrostatic and critical stresses respectively.
The (Beremin 1981) criterion is also a stress-based criterion. This criterion has the advantage
of having a stress concentration factor (ks) that considers the shape of the particle and is expressed
as follows:
ππΌπππ₯ + ππ (πππ β π0) = ππΆ (2-20)
where ππΌπππ₯, π0 are the maximum principal stress in the loading direction and the equivalent stress
within the matrix respectively.
The (Chu and Needleman 1980) criterion is a strain based criterion. It proposes that void
nucleation appears once a critical strain is reached. This criterion is used in the well-known
Gurson- Tvergaard-Needleman (GTN) model.
Page 34
19
It is believed that void nucleation is a continuous process due to inhomogeneity and local
stress concentrations within the material. Voids are found to nucleate at faster rates when
secondary phase particles, impurities, and large inclusions are present. It has also been observed
that voids favor decohesion at the soft matrix, while the hard matrix material usually leads to
particle fracture (Babout, Brechet et al. 2004).
2.4.2 Void Growth
It has been established that by (McClintock 1968) that hydrostatic stress, also known as
stress triaxiality is the crucial parameter in controlling void dilatation. In most cases, a direct
relation between stress triaxiality and void growth rates can be established. Recent studies by
(Weck 2007, Maire, Bouaziz et al. 2008, Landron 2011) using X-ray tomography have also
concluded the importance of the stress triaxiality in enhancing void growth.
Voids are highly dispersed within a matrix which makes them hard to study and analyze.
Therefore, (Landron 2011) considered the average diameter of twenty of the largest voids within
a population as a representation of the growth of a single void. She concluded that the initial
microstructures of the specimen would affect the void growth rates. She observed that dual phased
steel specimens with higher yield strength experienced faster growth rates.
Many computational studies were performed (unit cell modeling) to understand the
mechanism of void dilations (Hill 1984, Becker, Needleman et al. 1988, Koplik and Needleman
1988, Kuna and Sun 1996). Direct relations between initial void sizes and hydrostatic stresses have
been established. It was also concluded that large voids were observed at high-stress triaxiality.
Page 35
20
2.4.3 Void Coalescence
Adjacent voids tend to coalesce after the growth has reached its critical state fc. According
to (Koplik and Needleman 1988) this critical value varies slowly with hydrostatic stress and matrix
strain hardening. However, initial void volume fraction has a strong influence on this critical value.
The phenomena of voids joining are usually seen near the fracture of the specimen. The mechanism
of coalescence has not been studied extensively, due to its speed of occurrence and difficulty of
distinguishing. It is believed that coalescence occurs at three stages, namely internal necking
(Thompson 1987), shear localization (Weck 2007) and necklace coalescence (Benzerga 2000).
The internal necking can be defined as the breakage of the ligament between 2 voids, and it is the
most encountered type. (Weck 2007) has illustrated that localization is the primary phenomenon
that is usually companioned with void coalescence. Coalescence has many factors that influence
its behaviors. These factors include; initial void volume fraction, void spacing, and ligament
length, void shapes, and void distributions effects. Many studies on these different effects have
been carried on throughout the last decade. These studies resulted in different model criteria to
describe such behaviors. Experiments involving different notch sizes performed by (Weck 2007)
illustrated that a drastic decrease in coalescence is observed with an increase in stress triaxiality.
Therefore, Weck proposed a modified version of the Brown and Embury model that takes into
account void ligament length in modeling the coalescence behavior under different stress states.
2.4.4 Void Closure and Healing
Under compressive loadings, voids tend to close as illustrated in Figure 2-3. (StΓ₯hlberg 1986)
demonstrated that small voids are more accessible to remove than large ones. It was observed that
Page 36
21
void volume fraction may increase at the early stages due to enormous stresses, but will finally
decrease resulting in void closure. The principal recognized parameters that influence the closure
process are temperature, pressure, compressive stresses and healing time (Afshan 2013). It was
observed by (Wang, Thomson et al. 1996) that the closure rate needs a specified level of stress
triaxiality to achieve the closure. It was also observed that high temperatures enhance the process
enormously. The concept of void closure and healing is applicable in metal joining (Wei, Han et
al. 2004, Wang, Huang et al. 2007, Afshan 2013).
Figure 2-3: Void closure steps forming a crack (Zhang, Cui et al. 2009)
2.4.5 Void Size and Shape Effects
It has been established by many scientists that void size dramatically influences the growth
rate (Fond, Lobbrecht et al. 1996, Fleck 1997, Zhang, Bai et al. 1999, Gao, Wang et al. 2005). In
general, larger voids tend to have a faster growth rate than small voids. The intrinsic material
length, when considered, accounts for the void size effect (Liu, Qiu et al. 2003). (Wen, Huang et
al. 2005) extended the Gurson model based on the Taylor dislocation model to include void size
Page 37
22
effects. It was noted that a direct relation could be established between void size and initial void
volume fraction, where void size effects could be significant at some stage.
Page 38
23
CHAPTER 3: EXPERIMENTAL METHODS AND SPECIMEN
DESIGN
The material being investigated in this paper is AISI 4340 steel Alloy. The mechanical
properties of this material were captured using MTS machine. Tests were carried out at room
temperature under different loading conditions. All deformations were measured and recorded by
a Digital Image Correlation (DIC) system.
3.1 Material Specification and Heat-treatment
AISI 4340 steel is a medium carbon, low alloy steel known for its toughness and strength in
relatively large sections(ASTM-STEEL). It is also one of the existing types of nickel-chromium-
molybdenum steel materials. This material has excellent shock and impact resistance as well as
wear and abrasion resistance in the hardened condition (ASTM-STEEL). The properties of this
material offer good ductility in the annealed condition, allowing it to be bent or formed. This
material can also be machined by many conventional methods. For its chemical composition,
please refer to Table 3-1.
Table 3-1: Chemical composition in %wt. of Steel 4340 being tested
C Si Mn P S Cr Mo Ni Cu
0.4 0.25 0.73 0.004 0.0008 0.63 0.25 1.76 0.13
Page 39
24
The heat-treatment of this material is performed by, Braddock Metallurgical Inc.-Daytona Beach,
which is an accredited heat-treatment facility in Florida.
The heat-treatment process for the 32HRC was summarized as follows:
a) Parts were preheated at 1000 ΛF (Β±10) for 60 minutes in furnace #643 then loaded into a
hardening furnace hot.
b) Parts were hardened in a furnace #630 at 1525 ΛF (Β±25) for 60 minutes then oil quenched.
c) Parts were tempered in a furnace #632 at 1100 ΛF (Β±15) for 120 minutes, and air cooled to
room temperature.
d) All parts were processed in a mercury-free environment
The heat-treatment process for the 39HRC was summarized as follows:
a) Parts were preheated at 1200 ΛF (Β±25) for 120 minutes
b) Parts were hardened at 1525 ΛF (Β±25) for 60 minutes then oil quenched.
c) Parts were tempered in a furnace at 950 ΛF (Β±15) for 120 minutes and air cooled to room
temperature.
Results of heat-treatment indicate that the hardness achieved correspond to 32HRC and
39HRC respectively. All heat-treatment was conducted within the requirements of AMS
2750E.
Page 40
25
3.2 Representation of Stress State
The state of stress of isotropic material is commonly defined in Cartesian
coordinates (1 ,2 ,3 ). The plasticity, pressure and Lode angle model (PPL) model convert
stress from a Cartesian representation to a spherical representation. The three stress invariants (p,
q, r) can be defined as follows:
p = βm = 1
3 tr([]) = β
1
3( 1 + 2 + 3 )
(3-1)
q = = β3π½2 = β3
2 [S] βΆ [S] = β
1
2[(1 β 2)
2 + (2 β 3)2 + (3 β 1)
2]
(3-2)
r = (9
2[S] β [S]: [S])
13β
= [27
2det([S])]
13β
= [27
2(1 β m)(2 β m)(3 β m)]
13β
(3-3)
[S] = [ ] + p[I] (3-4)
The parameters [S] and [I] respectively denote the deviatoric stress tensor and the identity tensor.
It is assumed that the principal stresses are represented by 1 β₯ 2 β₯ 3. For convenience stress
triaxiality () is defined as:
= β p
q=
m
(3-5)
Page 41
26
The parameter has been widely used in many ductile fracture formulations (McClintock
1968, Rice and Tracey 1969, Hancock and Mackenzie 1976, Bao and Wierzbicki 2004). However,
the incorporation of Lode parameter is neglected in many of the previous models. The geometrical
meaning of Lode angle is illustrated in Figure 3-1.
Figure 3-1: Principal stresses in spherical coordinate representation (Bai and Wierzbicki 2008)
The Lode parameter is related to the third deviatoric stress invariant through the following
(Malvern 1969) equations:
= (r
q)3
= cos(3) (3-6)
Furthermore, the Lode angle parameter οΏ½οΏ½ is normalized and is expressed as:
Page 42
27
οΏ½οΏ½ = 1 β 6
= 1 β
2
π πππ β1( )
(3-7)
The state of stress of the isotropic material is usually defined in Cartesian
coordinate (1 ,2 ,3 ). The Cartesian coordinates can be simply transformed into either
cylindrical or spherical coordinate by appropriate mapping methods. From Figure 3-1, the vector
ππβ² is perpendicular to the plane that is located at the origin O, and every point on it represents
a hydrostatic stress state. Point (P) represents an arbitrary point that is located on the πβ² plane. To
describe the state of stress of point (P) in spherical coordinates the vector ππ must be determined.
By vector addition analogy, ππ is described by the addition of the two vectors ππβ² and πβ²P
(Bai and Wierzbicki 2008).
3.3 Design of Specimen Geometries
This study investigated 29 specimens of different shapes and at different tempering
conditions. There were 13 round bars (axial symmetric tension), 13 flat grooved (plane stain)
samples and three small round cylinders (upsetting test). Six round bars and six flat grooved
specimens were heat treated with 32 Rockwell Hardness (32 HRC). Four other round bars and four
flat grooved specimens were heat treated with 39 Rockwell Hardness (39 HRC). The hardness for
the initially untreated specimens was 16 Rockwell Hardness (16 HRC). The round bars are named
as follows: smooth bar (SB), large notch (LN) sharp notch (SN) and a small compressive cylinder
(SCC). The flat grooved were named as follows; large flat grooved (LFG), medium flat grooved
(MFG) and small flat grooved (SFG). Some of these specimens are illustrated in Figure 3-2 and
Figure 3-3. The design of different shapes enables to study the effects of varying stress states.
Page 43
28
These specimens are summarized in Table 3-2 and Table 3-3. Detail drawings are shown in Figure
3-4 through Figure 3-6.
Table 3-2: Round bar specimenβs names and dimensions (mm)
Specimen Name Symbol Notch
Radius
Minimum
Diameter
16 HRC Smooth Bar 16 SB N/A 5
32 HRC Heat-treated Smooth Bar 32 SB N/A 5
39 HRC Heat-treated Smooth Bar 39 SB N/A 5
16 HRC Small Compressive Cylinder 16 SCC N/A 7
16 HRC Large Notched Bar 16 LN 7.5 5
32 HRC Heat-treated Large Notched Bar 32 LN 7.5 5
16 HRC Sharp Notched Bar 16 SN 3.5 5
32 HRC Heat-treated Sharp Notched Bar 32 SN 3.5 5
39 HRC Heat-treated Sharp Notched Bar 39 SN 3.5 5
Page 44
29
Table 3-3: Flat grooved specimenβs names and dimensions (mm)
Specimen Name Symbol Notch
Radius
Minimum
Thickness
16 HRC Large Flat Grooved 16 LFG 40.625 1
32 HRC Heat-treated Large Flat Grooved 32 LFG 40.625 1
39 HRC Heat-treated Large Flat Grooved 39 LFG 40.625 1
16 HRC Medium Flat Grooved 16 MFG 10 1
32 HRC Heat Treated Medium Flat Grooved 32 MFG 10 1
16 HRC Small Flat Grooved 16 SFG 1.575 1
32 HRC Heat-treated Small Flat Grooved 32 SFG 1.575 1
39 HRC Heat-treated Small Flat Grooved 39 SFG 1.575 1
Page 45
30
Figure 3-2: Round bars and flat grooved specimens in the unheated case (16 HRC)
Figure 3-3: samples of other round bar specimens with different heat-treatment conditions
Page 46
31
Figure 3-4: Drawings of the small compressive cylinder
Figure 3-5: Drawings of round bars (SB, LN, and SN)
Page 47
32
Figure 3-6: Drawings of flat grooved specimens (FLG, FMG, and FSG)
3.4 Theoretical Analysis of Stress States
In theory, each specimen with different geometry shall experience different stress
state ( , ΞΈ). Fracture strain can be estimated using the area reduction method of the fractured
specimen. Lode angle parameter ( ) and stress triaxiality () can be determined using analytical
solutions. Please refer to Ref. (Bai and Wierzbicki 2008, Algarni, Bai et al. 2015) for more details.
The smooth bar specimen (SB) will be used to calibrate the necessary mechanical properties
including the strain hardening curve. The process of analyzing SB tensile test data provided by the
MTS machine and the DIC software are as follows: First, load-displacement curves are recorded
and analyzed. The curves are then converted to the engineering stress-strain curves (E, ΖE). The
engineering stress-strain curves are then converted to true stress-strain curves up to the necking
Page 48
33
initiation point. Once necking is observed, a Swift power function is introduced to describe strain
hardening beyond necking. Several FEA simulations are run to calibrate the parameters of the
swift power function until an excellent correlation between FEA and experimental load-
displacement results is achieved. Once FEA and experimental data match, an approximate true
stress-strain is established. Please refer to equations (3-8) through (3-10) .
π‘ = E (1 + ΖE ) (3-8)
Ζπ‘ = ln (1 + ΖE) (3-9)
(Ζππ) = K ( Ζ0 + Ζππ)π (3-10)
where πΈ , ΖπΈ , π‘ , Ζπ‘ , Ζ0, Ζππ , K, and n, are engineering stress-strain, true stress-strain, first yield
strain, strength index, and strain hardening exponent, consecutively.
The cylindrical compression specimens enable the determination of whether the shape of the
yield surface is symmetric or asymmetric. It has been proven that both the yield and fracture loci
are not necessarily symmetric (Bai and Wierzbicki 2008, Algarni, Bai et al. 2015, Bai and
Wierzbicki 2015, Bai Massachusetts Institute of Technology ;2008). Please refer to Figure 3-4
through Figure 3-6 for the detailed design of the specimenβs geometry.
The notched specimens are designed to investigate the effects of stress triaxiality () on
materialβs behavior. The deeper the notch is, the higher the stress triaxiality (). From analytical
solutions, the round bar specimens (axial symmetric tension) will exhibit a Lode angle parameter
of unity οΏ½οΏ½ = 1, while the plane strain specimens (Flat Grooved) and the small compressive
cylinder will exhibit Lode angle parameters of οΏ½οΏ½ = 0,β1 respectively. Thus, a comprehensive
Page 49
34
study of the effects of stress triaxiality () and Lode angle parameter dependence on materialβs
plasticity behavior is achieved. Please refer to Figure 3-5 for detailed drawings of the round
specimens.
The plane strain specimens have the same range of stress triaxialities () as the round bar
specimens, but the Lode angle parameter is ΞΈ = 0. See Figure 3-6 for specimen drawings. This
fact, in association with previous tests, will adequately determine the shape of the fracture locus.
Analytical solutions for plane strain specimens are illustrated below (Bai 2008, Algarni 2015).
Ζπ =2
β3 ln (
tot )
(3-11)
= m
=β3
3 [1 + 2 ln (1 +
π
2π
)]
(3-12)
xx = 2
β3 ln (
a2 + 2aR β π₯2
2aR)
(3-13)
yy =2
β3 [
1
2+ ln (
a2 + 2aR β x2
2aR)]
(3-14)
zz =2
β3 [1 + ln (
a2 + 2aR β x2
2aR)]
(3-15)
Page 50
35
3.5 Experimental Setup
Twenty-nine samples of different shapes and different heat-treatment were fabricated from
an AISI 4340 steel alloy square bar material. The testing took place at room temperature with
quasi-static loading conditions. All tests were carried out using an MTS testing machine that can
deliver up to 100 (KN). The selected loading rate was 0.15 (mm/s). The displacement of the
specimens was recorded by a camera and analyzed by DIC software. The period selected for all
experiments was one frame per second. The DIC requires a pattern spray on the specimen to enable
proper displacement depiction and detection. For this purpose, all specimens were painted using
an airbrush. An alternative method to measure a specimenβs displacement is by using an
extensometer. This method was not used here. Please refer to Figure 3-7 for test configuration
setup and to Figure 3-8 and Figure 3-9 for samples of fractured specimens.
Page 51
36
Figure 3-7: Test configuration setup
Page 52
37
Figure 3-8: Samples of round bar fractured specimens (cup and cone fractured surface)
Figure 3-9: Samples of flat grooved fractured specimens (slant fractured surface)
Page 53
38
CHAPTER 4: PLASTICITY AND FRACTURE MODELS CALIBRATION AND NUMERICAL
VERIFICATION
In this section, the Drucker and Prager and the Bai and Wierzbicki (PPL) models are
presented. The PPL and the Drucker and Prager models are both pressure dependent models.
However, the PPL model also incorporates the effects of the Lode parameter in a merit method
that considers different stress loading conditions.
4.1 The Drucker and Prager Plasticity Model
The Drucker and Prager (DP) model (Drucker and Prager 1952) is a pressure dependence
model that has been extensively used to describe pressure sensitive materials such as rocks and
concrete. This model and many of its extensions are built in many software packages including the
ABAQUS software.
The linear Drucker- Prager model (Simulia 2017) reads as follows:
πΉ = π‘ β π tan Ξ² β d = 0 (4-1)
π‘ =1
2 π [1 +
1
πΎβ (1 β
1
πΎ) (
π
π)3
] (4-2)
K is defined as the ratio of the yield stress in triaxial tension to that of triaxial compression, and K
is suggested to be in the range of 0.778 β€ πΎ β€ 1.0 to ensure the convexity of the yield surface. In
the case of uniaxial tension, the relationship between parameter π and yield stress is,
Page 54
39
π = (1 +1
3tanπ½)ππ‘
(4-3)
The angle Ξ² is often referred to as the friction angle and d as the cohesion of the material.
The plastic potential function G in the DP model is defined in accordance with the following
equation:
πΊ = π‘ β π tanπ (4-4)
It is worth noting that the original (DP) model is retrieved by Ξ²= π and πΎ = 1.
4.2 The PPL Model and Failure Criterion
4.2.1 Modification to the Original Bai and Wierzbicki Plasticity Model
The calibration of AISI 4340 steel was obtained by invoking the Bai and Wierzbicki (PPL)
plasticity model (Bai and Wierzbicki 2008, Algarni, Bai et al. 2015). This model grasps the effects
of plastic strain, pressure dependence, and Lode angle effects, on plasticity and fracture. The model
can be expressed as:
π¦ππ(Ζpl , , ) = (Ζππ)[1 β c (Ξ· β o)] [c
s + (cax β c
s) (π + 1
π)(β
m+1
m+ 1)] (4-5)
(Algarni 2015) recently introduced a new correction factor (π+1
π) to make the model more
user-friendly. Additionally, please note that the stress triaxiality term (Ξ·) in the original model is
modified to be (Ξ·), also known as modified stress triaxiality ratio, defined as:
Page 55
40
οΏ½οΏ½ =πππβ
(4-6)
where πβ = (Ζππ) is the yield stress at the reference loading of πβ. This modification ensures
the stability of the model at extreme loading conditions (i.e.,, when the q β 0 in Eq. (3-5)).
(Algarni 2015) claimed that the cs could be a function of the Lode angle parameter. A similar
claim could be issued regarding the c parameter (APPENDIX A). However, for calibration
proposes only, it is best that these two parameters (cs, c) are kept as constant parameters as
intended by the original model derivation (Bai and Wierzbicki 2008).
A Swift power function is introduced to describe the term (Ζpl) for metal plasticity as seen
in Eq. (3-10). The c is a coefficient that controls the effects of hydrostatic pressure on yielding
and needs to be calibrated. The cs, c
ax and π are coefficients that control the effects of Lode angle
on material plasticity. The and cax parameters are defined in Eq. (4-7) and (4-8) as follows:
= 6.4641 [sec( β /6) β 1] (4-7)
cax = {
ct for β₯ 0
cc for Λ 0
(4-8)
The terms ct and c
c are parameters denoting tension and compression respectively. It is worth
noting that the parameters ct = 1 and
o= 1 3β are uniquely defined if a smooth bar tension test
is used as a reference test (Bai and Wierzbicki 2008).
Most yield criteria in literature lack either pressure or Lode angle dependencies. Yield
criteria that include both pressure and Lode dependence are not easy to calibrate. The PPL model
Page 56
41
consists of both of the former dependencies and adopts a simplified calibration approach. The PPL
model is also very convenient for describing the state of stress at low, intermediate, and high-stress
triaxialities, consecutively. The parameters of the PPL plasticity model (π, ct , c
c, cs , π) can be
adjusted to match many of the well-known yield criteria, as seen in Table 4-1. Furthermore, the
PPL model predicts adequate results for specimens experiencing different stress states, which
makes it viable for industrial usage (i.e., all notched round bars and flat grooved specimens are
calibrated by a fixed set of PPL parameters).
Table 4-1: Achieving different yield criterions by adjusting the PPL parameters (Algarni 2017)
Yield Criterion π ππ π
π ππ¬ π
Von-Mises 0 0 0 0 0
Tresca 0 β3
2
1 1 +
Pressure-modified Tresca β 0 β3
2
1 1 +
Page 57
42
4.2.2 The Deviatoric Associated Flow Rule of the Modified PPL Model
The conventional, associated flow rule, can be expressed as
πνππππ= ππ
ππ
ππππ (4-9)
where the plastic flow is defined as:
π = π β οΏ½οΏ½(νπ) [1 β πΞ· (Ξ· β π0)] [πππ + (ππ
ππ₯ β πππ )(πΎ β
πΎπ+1
π + 1 )(1 +
1
π)]
(4-10)
Variable π is defined in Eq. (3-2) and ππ is the equivalent plastic strain increment.
The conventionally associated flow rule follows:
ππ
ππππ=ππ
ππππ+ οΏ½οΏ½(νπ) πΞ· [ππ
π + (ππππ₯ β ππ
π )(πΎ β πΎπ+1
π + 1 )(1 +
1
π)]
πΞ·
ππππ
β οΏ½οΏ½(νπ) [ 1 β πΞ· (Ξ· β π0)] (ππππ₯ β ππ
π )(1 +1
π)(1 β πΎπ )
ππΎ
ππππ
(4-11)
Where ππ
ππππ,ποΏ½οΏ½
ππππ and
ππΎ
ππππ are expressed as:
ππ
ππππ=
3
2 π π ππ
(4-12)
ποΏ½οΏ½
ππππ=
πΏππ3 πβ
(4-13)
ππΎ
ππππ = (
3β3
2 β β3)(
tan(π βπ6)
cos(π βπ6))
1
π sin (3π) (πΏππ3+cos (3π)
2π π ππ β
3
2π2π πππ ππ)
(4-14)
Page 58
43
To satisfy the assumption of plastic incompressibility, the term πΏππ
3 πβ in Eq. (4-13) is removed.
Thus, this paper uses deviatoric associativity rather than the conventional associativity as
illustrated in Eq. (4-15).
πνππππ= ππ {
3
2 π π ππ β οΏ½οΏ½(ν
π) [1 β πΞ·(Ξ· β π0)] (ππππ₯ β ππ
π )(1 +1
π)(1 β πΎπ )
Γ (3β3
2 β β3)(
tan(π βπ6)
cos(π βπ6))
1
π sin (3π) (πΏππ3+cos (3π)
2π π ππ
β3
2π2π πππ ππ)}
(4-15)
4.2.3 Modified Mohr-Coulomb Criterion Failure Criteria
To construct the fracture locus of a material, it is vital to correctly capture the fracture
initiation locations of the specimens that are under different stress states. This is obtained by
invoking the modified Mohr-Coulomb criterion (MMC). The MMC is expressed as (Bai and
Wierzbicki 2009):
Ζπ = {π΄
c2 [c
s +β3
2 β β3(cax β c
s) (sec (
6) β 1)] [β
1 + c12
3 cos (
6)
+ c1 ( + 1
3 sin (
6))]}
β1/π
(4-16)
Page 59
44
The parameters (π΄, π, πΆοΏ½οΏ½, πΆ
οΏ½οΏ½ , πΆ1, πΆ2) need to be calibrated. The parameters A and N are power
strain hardening parameters that can be found from the reference test (smooth bar). Other
parameters are then found by the inverse method (trial and error) until all simulated specimens
agree with their experimental values. MATLAB code was developed to obtain the initial
parameters of the MMC fracture locus (Algarni, Bai et al. 2015). It is assumed that fracture is
initiated when the damage indicator (D) defined in Eq. (4-17) reaches unity.
π· = π· (Ζππ) = β«πΖππ
Ζπ( , )
Ζππ
0
(4-17)
4.3 Calibration Procedure
A series of tests were conducted to determine the material parameters. The first test is the
smooth bar, which is taken as the reference test. The smooth bar test uniquely identifies the power
function parameters (πΎ, Ζ0, π) pre-necking described in Eq. (3-10) and Table 4-2.
Figure 4-1 indicates the true pre-necking stress-strain curve represented using the solid
curve, while post-necking is represented by dashed curves with diamond symbols. The post-
necking points were obtained in such a way that experimental data correlated with the simulated
data by the inverse method (or known as trial-and-error method). The DIC enables continuous
tracking of the minimum cross-sectional area of the specimen at all times. Knowing the exact
dimensions allows the utilizing Bridgmanβs stress-strain solutions. This method was also used to
verify the post necking behavior of the specimen. Please refer to APPENDIX B for detailed
derivations.
Page 60
45
The round notched bars and the flat grooved specimens of different sizes enable the study of
the modified stress triaxiality (Ξ·), and the Lode angle effects on material plasticity in a sequential
manner. The small compressive cylinder is used to study the material plasticity behaviors in the
negative stress triaxiality region. A friction coefficient of 0.05 is assumed for simulations involving
small compressive cylinder.
Table 4-2: List of material fundamental plasticity properties used in the FEA
Material Hardness Elastic
Modulus
E (GPa)
Poisson ratio
Ξ½ (-)
y
(MPa)
K
(MPa)
Ζπ
(-)
n
(-)
16 HRC 208 0.29 535.58 1151 0.0040 0.140
32 HRC 208 0.29 946.8 1387 0.0100 0.088
39 HRC 208 0.29 1122.6 1638 0.0112 0.086
Page 61
46
Figure 4-1: Comparison of stress-strain curves of heated (32 Rockwell) and unheated (16
Rockwell) specimens
4.3.1 Calibration of the Drucker and Prager Plasticity Model
The linear Drucker and Prager (DP) plasticity model built in ABAQUS has three constants
that needed calibration (Ξ², K, and Ο). As mentioned above, the smooth round bar provides the true
stress-strain of the material. After the determination of the true stress-strain data, a parametric
study of the three constants was conducted using the software ISIGHT provided by SIMULIA,
until optimum solutions were achieved. The calibrated plasticity parameters using the DP model
are listed in Table 4-3.
Page 62
47
Table 4-3: List of material parameters used in the DP plasticity model
AISI 4340 steel hardness Ξ² K Ο
16 HRC 10Β° 1 0Β°
32 HRC 5Β° 1 0Β°
39 HRC 5Β° 1 0Β°
4.3.2 Calibration of the PPL Plasticity Model
The second group of tests comprises round bars with different notch sizes. The introduction
of different notch sizes enables the study of the effects of stress triaxiality (Ξ·) on material plasticity,
and the determination of cΞ· and m variables. The third group of tests comprises flat grooved
specimens of different sizes. This group enables the study of Lode angle effects on material
plasticity and the determination of cΞΈs . The final test is the small cylindrical upsetting test, which
determines the cc variable. The calibrated plasticity parameters using the PPL model are listed in
Table 4-4. It is important to note that the small compressive cylinders were not heat-treated because
their strength would exceed our load cell and fixture capabilities. Therefore, the cc values for the
heat-treated cases were assumed to be unity.
The MMC model was used for predicting the initiation of the fracture. Given knowledge of
the true stress-strain curve for the smooth bar, the power hardening parameters (A and n) are fitted
based on the entire true stress-strain curves. Table 4-5 shows the MMC parameters for the three
types of material hardness. The stress triaxiality (Ξ·), Lode angle parameter ΞΈ, and fracture strain
(Ζπ) were estimated based on the theoretical formulas. The MMC model was then calibrated, and
results are summarized in Table 4-5.
Page 63
48
Table 4-4: List of material parameters used in the PPL plasticity model.
Material description π π¨ π
π ππ¬ m
16 HRC 0.2 13β 0.75 0.93 0.75
32 HRC 0.2 13β 1.0 1.03 0.75
39 HRC 0.05 13β 1.0 0.93 0.75
Table 4-5: List of material fracture parameters using MMC model
Material description A n C1 C2 ποΏ½οΏ½
16 HRC 1151 0.14 0.157 647 1.1
32 HRC 1383 0.0806 0.085 732 0.99
39 HRC 1500 0.0526 0.055 777 0.96
Page 64
49
4.4 Comparison of Experimental and Numerical Simulations Results
This section presents finite element simulation results compared side by side with
experimental results. The PPL was run using Abaqus/Explicit with a VUMAT subroutine. The
round specimens (SB, LN, and SN) were simulated using a quarter of the model with axisymmetric
elements (CAX4R) type owing to symmetry. Conversely, the flat grooved specimens (LFG, MFG,
and SFG) were constructed using one-eighth of the model with 3D solid elements (C3D8R) type
owing to symmetry. The small compressive cylinder (SCC) was constructed using 3D solid
elements (C3D8R) type as seen in Figure 4-2.
Page 65
50
Figure 4-2: Samples of meshes used in Abaqus for simulations
Page 66
51
4.4.1 Plastic Strength and Flow of AISI 4340 Steel with Different Hardness
A direct comparison of FEA and experimental results for the necking configuration observed
immediately before the fracture is illustrated in Figure 4-3 through Figure 4-9. Diffuse necking
for flat grooved specimens is indicated by circles.
Each fractured specimen was measured at various locations, and an average fractured
diameter was then determined. Table 4-6 and Table 4-7 compare the measurements of specimens
before and after tests with FE results. The FE fracture results were within a 5% of the error margin,
compared to the experimental results.
Figure 4-3:Comparison of plastic deformations between test and simulations for un-heated (16
HRC) round specimens
Page 67
52
Figure 4-4:Comparison of plastic deformations between test and simulations for heat treated
(32 HRC) round specimens
Page 68
53
Figure 4-5: Comparison of plastic deformations between test and simulations for heat treated
(39 HRC) round specimens
Page 69
54
Figure 4-6: Comparison of plastic deformations between test and simulations for un-heated
(16 HRC) flat grooved specimens
Figure 4-7: Comparison of plastic deformations between test and simulations for
heat treated (32 HRC) flat grooved specimens
Figure 4-8:Comparison of plastic deformations between test and simulations for
heat treated (39 HRC) flat grooved specimens
Page 70
55
Figure 4-9: Comparison of plastic deformations between test and simulations for un-heated
(16 Rockwell) small compressive cylinder
Page 71
56
Table 4-6: Dimensions of round bar specimens (axially symmetric) before and after the test.
Specimen
Name
Initial
diameter
(mm)
Test fractured
diameter (mm)
FEA
fractured
diameter
(mm)
Actual %
Area
reduction
%
Error
16 SB 5 3.42 3.25 31.6 4.97
16 LN 5 4.02 3.86 19.6 3.98
16 SN 5 3.65 3.55 27 2.74
32 SB 5 3.5 3.34 30 4.57
32 LN 5 3.88 3.67 22.4 5.41
32 SN 5 4.07 3.95 18.6 2.95
39 SB 5 3.54 3.48 29.2 1.69
39 SN 5 4.19 4.04 16.2 3.58
Table 4-7: Dimensions of flat grooved (plane strain) before and after the test.
Specimen
Name
Initial
thickness (mm)
Test fractured
diameter (mm)
FEA
fractured
diameter
(mm)
Actual %
Area
reduction
%
Error
16 LFG 1 0.87 0.83 13 4.60
16 MFG 1 0.90 0.86 10 4.44
16 SFG 1 0.93 0.89 7 4.30
32 LFG 1 0.88 0.85 12 3.41
32 MFG 1 0.90 0.87 10 3.33
32 SFG 1 0.92 0.89 8 3.26
39 LFG 1 0.92 0.88 8 4.35
39 SFG 1 0.94 0.90 6 4.26
Page 72
57
The results elicited by the Drucker and Prager model are in good agreement with the
experimental results when compared to the J2 results, as seen in Figure 4-10 through Figure 4-12.
The dashed curves and the dotted curves indicate the results corresponding to the DP and J2
simulations, respectively. The solid curves show the experimental results. Noteworthy is the fact
that Please note that the cross marks in all the simulations (Figure 4-10 through Figure 4-15)
indicate material failure by separation. The main advantage of the DP model is its ease of
calibration. However, the DP model cannot still accurately model the 16 HRC flat grooved
specimens seen in Figure 4-10. This is owed to the Lode angle dependence which is not accounted
for in the current DP model. The DP model has an asymmetric yield surface. Therefore, using the
tensile hardening curve data for compressive loading paths may lead to inadequate results for some
materials (Figure 4-16). The goal here is to use one hardening curve to capture all complex loading
conditions with acceptable accuracy (i.e., use a plasticity model that can capture the effects of
different loading conditions). The modified PPL model is designed to achieve the premise.
Good agreement with the experimental results is achieved using the modified PPL model.
This is well illustrated in Figure 4-13 through Figure 4-15. The dashed curves and the dotted curves
indicate the results elicited by the PPL and J2 simulations, respectively. The solid curves indicate
experimental results.
A proportional relation between hardness and strength can be established. It is observed that
increasing hardness will increase the strength and lower the ductility. Hence, an inverse
relationship between hardness and ductility is observed.
Page 73
58
Figure 4-10: Force-displacement curves of AISI 4340 steel (16 HRC) using DP model
Page 74
59
Figure 4-11: Force-displacement curves of AISI 4340 steel (32 HRC) using DP model
Page 75
60
Figure 4-12: Force-displacement curves of AISI 4340 steel (39 HRC) using DP model
Page 76
61
Figure 4-13: Force-displacement curves of AISI 4340 steel (16 HRC) using the PPL model
Page 77
62
Figure 4-14: Force-displacement curves of AISI 4340 steel (32 HRC) using the PPL model
Page 78
63
Figure 4-15: Force-displacement curves of AISI 4340 steel (39 HRC) using PPL model
Page 79
64
Figure 4-16: Comparison of true stress-strain curves from experimental and FEA results for
untreated AISI 4340 steel small compressive cylinder (16 HRC)
Page 80
65
4.4.2 Fracture Surface Determination
As previously mentioned, the axial symmetric round bars in tension determine the fracture
locus at = 1 and the flat grooved (plane strain) determines the locus at = 0. The small
compressive cylinder can give fracture data at = β1 but no fracture was observed. Figure 4-17
shows the 2D fracture locus of the axial symmetric tension = 1 and the plane strain in tension at
= 0, which shows the effects of stress triaxiality and Lode angle parameters. The calibrated 3D
fracture loci of different tempering conditions are shown in Figure 4-18 through Figure 4-20.
It is found that the heat-treated specimens (32 HRC and 39 HRC) have some very similar
fracture surfaces to that of the unheated specimens (16 HRC) with some shift downwards
indicating loss of ductility due to heat-treatment effects.
Page 81
66
Figure 4-17: Calibrated 2D fracture locus of axial symmetric tension (round bar
specimens) at = 1 and plane strain tension (flat grooved specimens) at = 0
Page 82
67
Figure 4-18: Calibrated 3D fracture locus of unheated AISI 4340 steel (16 HRC).
Page 83
68
Figure 4-19: Calibrated 3D fracture locus of heat-treated AISI 4340 steel (32 HRC).
Page 84
69
Figure 4-20: Calibrated 3D fracture locus of heat-treated AISI 4340 steel (39 HRC).
Page 85
70
4.5 Results Discussion of the PPL Model Coupled With MMC Fracture Criteria
The PPL model and MMC fracture criterion were used to study the mechanical properties of
AISI 4340 steel samples. Twenty-nine different specimens of different shapes and heat-treatments
were tested and were used to calibrate the models. It was observed that this material exhibited both
stress triaxiality and Lode angle dependence behaviors. It is also worth noting that the shape of the
yield surface of this material changed with heat-treatment effects. The PPL model for the unheated
case (16 HRC) was contained within the Tresca and von Mises yield surfaces. The results of the
heat treated case (32 HRC) exhibited a yield surface shape change and exceeded the von Mises
envelope. Further, heat-treatment (39 HRC) presented similar effects to those observed for the
unheated sample (16 HRC). This is well demonstrated in Figure 4-21, were the shapes of the initial
yield surfaces at the Ο-plane are plotted. This confirms previous studies (Tanaka and Spretnak
1973, Hickey and Anctil 1985, Podder, Mondal et al. 2012, Sirin, Sirin et al. 2013) showing that
heat-treatments not only affects mechanical properties but also changes the yield surface shape.
The 3D fracture loci of both the heat-treated and untreated samples revealed similarities. Note
that the compressive heat-treated cylinder test was not conducted because it was beyond our
machine capabilities. Therefore, the compression behaviors for the heat treated specimens (32 and
39 HRC) were assumed to be identical to those of the untreated specimens (16 HRC) for simplicity.
It is found that the heat-treatment increased strength and sacrificed ductility. These
mechanical differences are attributed to different microstructural void arrangements and void
sizes, as discussed in (Lee and Su 1999).
Page 86
71
Figure 4-21: Shapes of the initial yield surfaces at the Ο plane for AISI 4340 steel with different
hardness.
Page 87
72
CHAPTER 5: EXTENSION AND APPLICATION OF MICROMECHANICS BASED GURSON TYPE
MODEL
5.1 Model Framework and Previous Studies
Throughout the last few decades, many ductile fracture models have been proposed. Some
of these models can be classified into micromechanical or damage based models. The ductile
fracture can be defined as a process that involves growth, nucleation, and coalescence of
microvoids (Wilsdorf 1975). The mechanisms of void behavior were first identified by Tipper
(Tipper 1949). Studies on the micromechanics level of isolated voids in a plastically deforming
solid are usually attributed to the works of McClintock who analyzed circular cylindrical voids
(McClintock 1968). Conversely, Rice and Tracey analyzed spherical voids (Rice JR 1969). Their
analysis can overestimate the strain at which coalescence occurs due to internal necking between
cavities interrupted by localized shear (Tvergaard and Needleman 1984). Many studies indicate
that coalescence by localized shear occurs when the void spacing is a constant of order unity times
the void length (Rogers 1960, Cox and Low 1974, Hancock and Mackenzie 1976, Knott and Green
1976). Theories of microvoid growth using notch round bar specimens were verified by Hancock
and Mackenzie (Hancock and Mackenzie 1976) , Hancock and Brown (Hancock and Brown 1983).
Further developments of previous models were successfully carried out by (Gurson 1977)
in which he proposed a yield function that links the effects of preexisting voids along with matrix
material. (Chu and Needleman 1980, Tvergaard 1981, Tvergaard 1982, LauNielsen and Tvergaard
2009) modified Gursonβs work and included some constant parameters that control the plastic flow
potential. Nevertheless, a new void nucleation mechanism was developed in addition to void
Page 88
73
growth. This model is often referred to as GTN model or porous metal plasticity model and is
expressed as the following:
= (
π¦)
2
+ 2π1πβ cosh (
β3π2p
2π¦)β(1 + π3π
β2) = 0 (5-1)
Where πβ is defined as the following:
{
π ππ π β€ ππ
ππ +ποΏ½οΏ½ β ππππ β ππ
(π β ππ) ππ ππ < π < ππ
ποΏ½οΏ½ ππ π β₯ ππ
ποΏ½οΏ½ =π1 +βπ1
2 + π3π3
(5-2)
π = οΏ½οΏ½ππππ€π‘β + π ππ’πππππ‘πππ (5-3)
where ποΏ½οΏ½πππ€π‘β is based on the conservation of mass and is expressed in terms of void volume
fraction:
οΏ½οΏ½ππππ€π‘β = (1 β π)νππ: πΌ (5-4)
The π ππ’πππππ‘πππ is given as:
π ππ’πππππ‘πππ =ππ
π π β2π ππ₯π [
β1
2(Ξ΅πππβ Ξ΅ππ π
)
2
] Ξ΅πππ
(5-5)
where π¦ is the yield stress of the matrix material and p is the pressure. The GTN model has a total
of 9 constant parameters that need calibration. These constants include (q1, q2, q3), which are
Page 89
74
parameters in the plastic flow potential to account for interactions between cavities added by
Tvergaard (Tvergaard 1981). π0 is the initial void volume fraction. ππ is a critical volume void
fraction that when reached, the materialβs strength starts to decay more rapidly (i.e.,, softening
region). ππ is the value of the void volume fraction at which there is a complete loss of strength. νπ
is the nucleation strain mean value. π π is the standard deviation, and finally, ππ is the volume
fraction of the nucleated voids. The GTN model defines void volume fraction as the summation
of void growth and void nucleation terms as shown in Eq. (5-3). However, when using high x-ray
tomography resolutions; voids that appear to nucleate are nothing more than small voids expanding
(Landron 2011). Please note that void nucleation effects are often neglected (Gao, Faleskog et al.
1998, Zhang, Thaulow et al. 2000, Xue, Pontin et al. 2010).
It has been argued that the critical volume fraction (fc) is not constant, and is a function of
stress triaxiality (Zhang and Niemi 1994). The GTN model assumes that this (fc) is constant. In
general, (fc) is inversely related to stress triaxiality. In fact, (fc) is also dependent on the initially
assumed void volume fraction (Koplik and Needleman 1988, Tvergaard 1990). This constant
critical void volume fraction is usually found by trial and error of smooth axisymmetric round bar
specimen or by unit cell modeling. This indicates that this (fc) is not unique. Zhang using the
(Thomason 1990) criterion modified the GTN failure criterion to be independent of this fc . The
coalescence-failure criterion developed by Zhang is related to void geometries reaching an
unstable limit. The main advantage of this criterion is its ability to predict lower void volume
fractions at low-stress triaxiality compared to the traditional GTN. However, many assumptions
and approximations have been made. Furthermore, this criterion is limited to volume void fractions
Page 90
75
of 0.2. The applicability of this criterion for negative hydrostatic stress is also questionable and
needs to be addressed. It is worth noting that this criterion is also dependent on the initial void
volume fraction since it involves the dimensions of the void and the unit cell.
A main drawback to the GTN model is its unsuitability to model localization, and fracture
for low triaxiality and shear dominated deformations (Nahshon K 2008). Recent studies have also
shown the importance of Lode parameter as well as triaxiality to correctly capture the plasticity
and fracture behaviors of materials (Bao and Wierzbicki 2004, Bai Massachusetts Institute of
Technology ;2008).
Nahshon and Hutchinson (Nahshon K 2008) modified the GTN model to include the effects
of what was defined as βvoid shearβ which is another failure (damage) mechanism to describe the
microvoids shear localization with low triaxiality. This modification is purely phenomenological,
meaning that the parameter π is a damage parameter rather than representing the void volume
fraction. It is important to note that the βvoid shearβ is formulated such that it vanishes under an
axisymmetric stress state, i.e., retrieving the original GTN model. This new damage term βvoid
shearβ considers the Lode angle parameter effects.
(Nielsen and Tvergaard 2009) studied the effects of the shear modified Gurson model
developed by Nahshon and Hutchinson on damage development in friction stir welding tensile
specimens and concluded that the extra damage parameter has the expected effect of promoting
fracture at moderate levels of stress triaxiality .
Page 91
76
The initial void size has some great significance effects on void growth behavior as shown
in recent literature (Fond, Lobbrecht et al. 1996, Fleck 1997, Shu 1998, Huo, Zheng et al. 1999,
Zhang, Bai et al. 1999, Zhang and Hsia 2001). Huang (Wen, Hwang et al. 2005) extended the
Gurson model based on the Taylor dislocation model to account for void size effects by
considering the intrinsic material length π as a parameter representing voids. The extended Gurson
becomes handy when void sizes are thought to be significant especially at extreme loading
conditions. It is also important to note that different void shapes have been considered by various
researchers (Gologanu, Leblond et al. 1995, Pardoen and Hutchinson 2000). However, many of
these extensions have only considered voids under axisymmetric loading conditions and did not
address shear dominated deformations (Nahshon K 2008).
(Liu, Wong et al. 2016) studied the effects of Lode angle, stress triaxiality and initial void
volume fraction on voided unit cell from an energy perspective. The critical strain to onset void
collapse and coalescence was established by examining the energetics of the unit cell system
throughout the loading process. The primary goal of the study was to investigate the existence of
a transition zone between void collapse and coalescence which was proven.
To better understand the different mechanisms of void failure at the microscopic level,
utilization of x-ray tomography comes at a value. (Landron 2011) used x-ray tomography during
several fracture tests under different complex loading processes. Landron studied the effects of
void coalescence for the first time for an industrial. Landron concluded that the Thomason criterion
(Thomason 1990) is the most suitable criteria for coalescence for dual phase (DP) steels. Landron
also investigated the effects of void nucleation and concluded that bi-phase microstructure of DP
Page 92
77
steels has a much faster nucleation rate when compared to single phase steels. Therefore, the
mechanism of void nucleation is highly dependent on the materialβs microstructure. Lastly,
Landron proposed the idea of measuring the diameters of the largest twenty pores in the population
and taking the mean as a representative of the measurement of single voids for void growth
mechanism.
Instead of focusing only on the mathematical aspect developed by Landron, (Fansi 2013)
established a modified GTN model using the works of Landron and Ben Bettiab. The new model
is physically based on in-situ high-resolution X-ray tomography. Fansiβs model has three main
contributions.
Developed a new kinetic law for void nucleation predicting the evolution of void density
considering stress triaxiality effects.
The void growth model was based on the mean diameter of the population studied and
instead of modeling one void; many voids were modeled using equal spacing between
them.
The critical volume fraction ππ has been provided by measuring the mean distance
between two cavities provided by Landron.
This model was implemented in FORTRAN and linked to ABAQUS-EXPLICIT using
VUMAT subroutine. The model illustrates the effect and influence stress triaxiality has on void
growth and void density. Fansi observed that with different notch sizes the crack develops at a
different location. He noted that deep notches develop cracks near the bottom edge, while
specimens with larger notches tend to move the crack towards the center of the specimen.
Page 93
78
The main drawback of Fansiβs model is the need of high accurate X-ray tomography as well
as the increasing number parameters and constants that needs calibration. The main contributions
of Landron and Fansiβs work is a better understanding of the materialβs behavior at the microscopic
level.
A parametric study is essential to get a clear image of the influence of each GTN parameter
on material plasticity and fracture behavior. (Slimane, Bouchouicha et al. 2015) and his team
performed a parametric study on GTN model using three notched specimens of different sizes. The
study was to evaluate the effects of mesh size, ππ and q parameters influence. It was observed that
elastoplastic part of the material is mesh insensitive. However, the softening region is mesh
sensitive. A finer mesh results in earlier fracture due to the enhancement of void initiation. The
study also concluded that increasing the ππ parameter has no influence on the elastoplastic region.
Contrary, fracture occurrence is earlier for higher values. Lastly, the growth of qβs increases the
effects of void volume fraction, which causes a severe decrease in tensile strength. π1 parameter
variation to values greater than unity causes a loss of strength of GTN material, but this loss is not
significant when compared to the variation of parameter π2. Increasing π2 to values greater than
unity causes a premature drop in strength and fracture initiation (i.e., less ductility and rapid loss
of strength indicating fracture). The study concluded that q parameters are of significant influence.
Thus, many authors held to Tvergaard suggested values. Please refer to Table 5-1 for better
understanding.
Page 94
79
Table 5-1: Effects of varying GTN model parameters
Description Elastic
region
Plastic region Softening region (Fracture)
Mesh size No
effects
No effects Finer meshes cause earlier
fracture
ππ No
effects
No effects Increasing values cause
earlier fracture
ππ No
effects
Little effects near the softening
region
Increasing values cause
earlier fracture
ππ No
effects
Great effect. Values higher than
unity results in a premature plastic
region development (less ductility)
Increasing values cause
extreme premature fracture
(Kiran and Khandelwal 2014) reevaluated the capability of the GTN model in calibrating
ASTM A992 steels under different complex stress states. Stress triaxiality was investigated by
designing specimens of different notch sizes. Kiran illustrated the non-uniqueness of the material
parameters (νπ , ππ) indicating that ductile fracture initiation can be simulated as early nucleation
or delayed nucleation of large volume voids. Thus, a new physical criterion for determining a
unique void nucleation term (νπ) was established and recommended. The calibration process laid
out by Kiran is as following:
Page 95
80
1. Calibration of nucleation parameters (νπ, ππ) :
These parameters govern the plastic strain level at which damage is imitated and evolved.
Kiran has shown the non-uniqueness of these parameters (i.e., many combinations in the
domain will simulate the behavior of microvoid imitation and growth). Increasing νπ will
result in delayed nucleation. Besides, a high value of ππ indicates substantial nucleation
of voids occurring. Since there is little physical meaning to a unique set, Kiran proposed a
new method for calibrating νπ . He proposed that πΊπ΅ corresponds to the point of
divergence of experimental data from J2 plasticity see Figure 5-1. This technique adds
some physical meaning for νπ since the divergence of curves (J2 and experimental) indicate
significant void activity because J2 assumes no damage parameters and no voids. After
calibrating νπ , ππ is chosen in a merit way that best fits the experimental load-displacement
data.
2. Calibration of (ππ , ππ) :
These parameters govern the accelerated void volume growth and final volumetric failure.
Once void volume fraction reaches the critical ππ , an accelerated void growth occurs until
fracture. A parametric study of this unique set is required to get the best calibration. It is
worth noting that the (ππ, ππ) parameters are highly mesh sensitive.
3. Calibration of (π1, π2, π3, π0, π π):
(π1, π2, π3) are parameters that are often chosen as suggested by Tvergaard (i.e., π1 =
1.5, π2 = 1, π3 = 2.25). On the other hand, (π0, π π) values are often chosen to match
literature.
Page 96
81
Figure 5-1: Experimental load-displacement curve and the corresponding FEA results with J2
plasticity model (Kiran and Khandelwal 2014).
Page 97
82
In this paper, the plasticity and fracture behaviors of AISI steel 4340 samples under different
tempering conditions will be studied. The Gurson-Tvergaard Needleman (GTN) is used to describe
the material plasticity and fracture. Also, a modified Gurson-Tvergaard Needleman plasticity
model coupled with modified Mohr-Coulomb fracture model denoted as (GTN-MMC) is used for
the plasticity and fracture modeling consecutively. As mentioned previously, the Lode angle
parameter is crucial, especially when modeling shear dominated regions. Thus, the effects of this
parameter on matrix material should be included. A series of mechanical tests to fracture were
conducted using the MTS machine with Digital Image Correlation (DIC) analysis. The (GTN-
MMC) model was implemented into Abaqus/explicit by means of a user-defined subroutine
(VUMAT). The final analysis shows good correspondence between experimental and simulated
results.
5.2 Theoretical Improvements to the GTN Model
The Gurson-Tvergaard-Needleman (GTN) model neglects the effects of the third deviatoric
stress invariant (or Lode angle parameter) on material plasticity. It is found that the GTN model
does not imitate the experimental data for plane strain specimens for the 16 Rockwell Hardness
(HRC) case as will be shown in section 5.3. Therefore, the GTN model needs improvements to
account for Lode angle dependency on matrix material. It is important to note that the proposed
modified model is a decoupled model, unlike the coupled GTN model. This section attempts to
address the former issues by proposing the following:
Page 98
83
5.2.1 Modified GTN Plasticity Model with Lode Angle Matrix Material Dependence
The GTN matrix material is considered a homogeneous, incompressible, rigid plastic, von
Mises material. We propose the following yield criterion suggested by Bai and Wierzbicki (Bai
2008):
π¦,πππ‘ = (Ζππ) [cs + (c
ax β cs) (
π + 1
π)(β
m+1
m+ 1)]
(5-6)
Where, (Ζππ) can be described by many hardening laws such as Swift, Ludwik , Voce , Isotropic
hardening, etc.
The ππ¬, π
ππ± and π are constant coefficients that controls the effects of Lode angle on
material plasticity. The () parameter is the difference between Von Mises and Tresca in the
deviatoric stress plane. The and πππ± are expressed as follows:
= 6.4641 [sec( β /6) β 1] (5-7)
cax = {
ct for β₯ 0
cc for Λ 0
(5-8)
This yield criterion has gained much attention in the recent literature (Wierzbicki, Bao et al.
2005, Beese, Luo et al. 2010, Algarni 2015, Bai 2015, Jia and Bai 2016, Algarni, Choi et al. 2017).
It also enables easy calibration of plane strain specimens using the GTN model. Using the
associated flow rule, a new plastic flow potential is adapted as follows:
dΞ΅ππππ = ππ
ππ
πππ
(5-9)
Page 99
84
Where π and πͺ are defined as the following:
π = q β π¦ [cs + (c
ax β cs) (
π + 1
π)
β (βm+1
m+ 1)]β(1 + π3π
β2)β 2π1πβ cosh (
β3π2π
2π¦)
(5-10)
Please note that this π is not the void volume fraction.
q = = β3π½2 = β3
2 [S] βΆ [S] = β
1
2[(1 β 2)
2 + (2 β 3)2 + (3 β 1)
2]
(5-11)
5.2.2 Modified Evolution Rule of Microvoid Volume Void Fraction (π)
The works of (Landron 2011, Fansi 2013) prove the linkage between void nucleation and
stress triaxiality. Higher stress triaxiality results in more nucleation activity. The current GTN
model doesnβt account for such linkage. Since nucleation is usually activated once ππ reaches a
constant value, it is hard to control the softening and fracture regions for different stress triaxialities
using the same ππ for all cases. The proposed new volumetric void fraction is as follows:
π = π0 + (πππππ€π‘β + πππ’πππππ‘πππ) β πΉ1(, π) (5-12)
The F(, π) can take many forms. However, the proposed form is as following:
πΉ1(, π) = (1 β cp (β o)) (1 + π1(1 β )
π2) (5-13)
Page 100
85
By choosing cp = m1 = 0, the GTN void volume fraction is retrieved. Introducing such terms
enables better calibration of specimens under different stress states. o is chosen from the reference
test (i.e., o= 1 3β for the smooth bar test). If the smooth bar is selected as the reference test cp
will be chosen in such a way to give the best fit for other specimens under different stress triaxiality
state. The advantage of using the Lode angle parameter is its simplicity. = 1,β1 for axial
symmetric tension and compression consecutively. = 0 for simple shear cases (Bai 2008). Since
= 0 for flat grooved specimens, π1 and π2 are chosen in such a way to obtain the best fit among
different grooves under different stress states. However, for simplicity and to attain the physical
meaning of π the original GTN void volume fraction is used for calibration (i.e., cp = m1 = m2 =
0).
5.2.3 Modified Criterion of Critical Microvoid
Volume Fraction to Fracture (ππ) Depending on Different Stress States
To obtain the best fit possible for different specimens under different loading conditions. It
was observed that neither ππ nor ππ is constant under different loading conditions (i.e.,, choosing a
fixed pair of ππ and ππ to calibrate specific loading conditions, for example, smooth bar, results in
under/over predicting fracture phenomena for different loading conditions). This behavior has also
been reported by (Kiran and Khandelwal 2014). Nahshon and Hutchinson have also noted that
many metallic materials have a fracture strain in shear that is well below fracture strain for
axisymmetric loadings (Nahshon K 2008). They suggested the addition of a new damage
parameter that could be called βvoid shearβ. However, this new damage parameter changes the
physical meaning of void volume fraction π in the sense that βvoid shearβ is a phenomenological
Page 101
86
term. Therefore, a new decoupled fracture criterion that keeps the physical meaning of the π
parameter is proposed and developed. This decoupled fracture criterion is similar to the extended
Mohr-Coulomb fracture criteria developed by Bai and Wierzbicki (Bai and Wierzbicki 2009).
The modified Mohr-Coulomb fracture criteria (MMC) has been gaining much attention in
recent literature (Bai and Wierzbicki 2009, Li, Luo et al. 2010, Luo and Wierzbicki 2010, Dunand
and Mohr 2011, Areias and Rabczuk 2013, Algarni 2015, Jia and Bai 2016, Jia, Ghazali et al.
2017, Jia, Qiao et al. 2017, Qiao, Liu et al. 2017). The Mohr-Coulomb (MC) is a physical simple
fracture criterion that has been widely used throughout literature (Zhao 2000, Chung, Ma et al.
2011, Algarni, Jia et al. 2015, Jia, Qiao et al. 2017). In this section a new decoupled fracture
criterion that links (π, , οΏ½οΏ½ ) is developed as follows:
The Gurson-Tvergaard-Needleman GTN model is expressed as follows:
= (
π¦,πππ‘)
2
+ 2π1π cosh (β3π2p
2π¦,πππ‘)β(1 + π3π
2) = 0 (5-14)
Noting that the πβ in the GTN has been replaced by π making it a complete decoupled model.
Solving Eq. (5-14) for :
= π¦,πππ‘β(1 + π3π2)β 2π1π cosh (
β3π2π
2π¦,πππ‘)
(5-15)
Expanding Eq. (5-15) and considering tiny volume void fractions π thus, π2 β
0
Page 102
87
β(1 + π3π2)β 2π1π cosh (
β3π2π
2π¦,πππ‘) = 1 +
1
2 [π3π
2 β 2π1π cosh (β3π2π
2π¦,πππ‘) ]
= [1β π1 π cosh (β3π2π
2π¦,πππ‘)]
(5-16)
Thus, the GTN is reduced to the following:
= π¦,πππ‘β(1 + π3π2)β 2π1π cosh (
β3π2π
2π¦,πππ‘)
β
π¦,πππ‘ [1β π1π cosh (β3π2π
2π¦,πππ‘)]
(5-17)
The Mohr-Coulomb in terms of ( , , π) is expressed as follows (Bai and Wierzbicki 2009):
= c2 [β(1 + c1
3) cos (
π
6β π) + c1 ( +
1
3 π ππ (
π
6β π))]
β1
(5-18)
π can be expressed as (Bai 2008) for simplicity:
π =π
6(1 β ΞΈ) (5-19)
Equating (5-17), (5-18) and (5-19) and denoting that the volume void fraction at the point of
fracture is ππ
Page 103
88
ππ = [π1cosh (β3π2π
2π¦,πππ‘)]
β1
[1
βc2
π¦,πππ‘[β(
1 + c1
3) cos (
πΞΈ
6) + c1 ( +
1
3 π ππ (
πΞΈ
6))]
β1
]
(5-20)
Where (Bai and Wierzbicki 2008);
π¦,πππ‘=π¦,π½2 β [οΏ½οΏ½π + (οΏ½οΏ½
ππ₯ β οΏ½οΏ½π ) (
β3
2 β β3[π ππ (
πΞΈ
6) β 1])]
(5-21)
We define and π2 following:
=π
π¦,π½2 (5-22)
c2 =c2
π¦,πππ‘
(5-23)
Thus (5-20) is reduced to the following:
ππ =
[
q1cosh
(
β3q2
2 [cs + (c
ax β cs) (
β3
2 β β3[sec (
ΟΞΈ6 )
β 1])])
] β1
[1
β c2 [β(1 + c1
3) cos (
ΟΞΈ
6) + c1 ( +
1
3 sin (
ΟΞΈ
6))]
β1
]
(5-24)
Page 104
89
This decoupled fracture criterion has a total of six parameters (π1, π2, c1, c2 , οΏ½οΏ½π , οΏ½οΏ½
ππ₯) that need to
be found.
5.3 Experimental Results of AISI 4340 Steel and Numerical Validation of
the Proposed (GTN-MMC) Model
In this section, an attempt to calibrate AISI 4340 steel of hardness 16 HRC, 32HRC and 39
HRC is carried out using the hardening curves shown in Figure 5-2 and the parameters shown at
Table 5-2 Note that the solid curves in Figure 5-2 indicate pre-necking behaviors, while the dashed
curves indicate the post necking. The post necking data was extrapolated using Swift hardening
law, and then data was adjusted through iteration to get the best FEA correlation with experiments
(trial and error method).
Figure 5-2: Hardening curves for 16 HRC, 32 HRC and 39 HRC consecutively used in GTN
model
Page 105
90
Table 5-2: GTN parameters used for 16 HRC, 32HRC and 39 HRC consecutively
Hardness ππ ππ ππ ππ ππ΅ πΊπ΅ ππ΅ ππ ππ
16 HRC 1.5 1 π12 0.0065 0.01 0.1 0.1 0.12 0.15
32 HRC 1.5 1 π12 0.0065 0.005 0.1 0.1 0.08 0.1
39HRC 1.5 1 π12 0.0065 0.005 0.1 0.1 0.08 0.08
The round specimens (SB, LN, and SN) were simulated using a quarter of the model with
an axisymmetric elements (CAX4R) type due to symmetry. While, the flat grooved specimens
(LFG, MFG, and SFG) were constructed using a 1 8β model with 3D solid elements (C3D8R) type
due to symmetry. The small compressive cylinder (SCC) was created using 3D solid elements
(C3D8R).
The q parameters are chosen as suggested by (Tvergaard 1981). π0 is chosen to resemble the
low porosity of this material. Other parameters are chosen based on the best fit for all cases and
within literature suggested values (Kiran and Khandelwal 2014). Note that the smooth bar is used
as the reference specimen for calibration purposes (i.e., fracture parameters (ππ , ππ) are chosen to
fit the smooth bar fracture location). It is clear from Figure 5-3, Figure 5-5 and Figure 5-6 that the
GTN model does not perceive the correct fracture location (denoted by cross mark) compared with
experimental tests (denoted by an asterisk). Choosing a unique set of (ππ , ππ) to fit the smooth bar
will lead to early fracture locations for round bar specimens under higher stress triaxiality states.
Picking the same fracture parameter set will also lead to extreme late fracture behaviors in-plane
Page 106
91
strain specimens when compared to the experimental tests. Therefore, the new decoupled fracture
criterion is implemented in the upcoming sections.
It can also be seen that round bar specimens give a reasonable plasticity fit using the GTN
model (dashed curves) compared with experimental (solid curves). However, the GTN model does
not provide adequate plasticity behavior for the cases of 16 HRC flat grooved (plane strain
specimens) as seen in Figure 5-3 nor the small compressive cylinder case as seen Figure 5-4. The
absence of Lode angle parameter is probably the reason for the inadequate plasticity fit achieved
for the flat grooved specimens. On the other hand, the existence of void nucleation terms in the
compressive region will affect the attainment of realistic results for the small compressive cylinder
case (SCC).
The previous observations are evidence that the current GTN model needs improvements to
include Lode angle parameter effects on material plasticity, especially for the 16 HRC flat grooved
(plane strain) specimens. However, the cases of 32 HRC (Figure 5-5) and 39 HRC (Figure 5-6)
seem to have less plasticity dependence on Lode angle, and cs of unity is therefore chosen. It is
worth noting that choosing cs to be unity retrieves the original GTN yield criterion.
Page 107
92
Figure 5-3: Force-displacement curves for (16 HRC) GTN model
Page 108
93
Figure 5-4: Stress-strain curve of the small compressive cylinder (16 HRC)
Page 109
94
Figure 5-5: Force-displacement curves for (32 HRC)
Page 110
95
Figure 5-6: Force-displacement curves for (39 HRC)
Page 111
96
5.3.1 Validation of Proposed Model (GTN-MMC)
In this section, an attempt to calibrate AISI 4340 steel of hardness 16 HRC, 32HRC and 39
HRC is carried out using the updated yield criterion Eq. (5-6) and decoupled fracture criterion Eq.
(5-24) (GTN-MMC).
5.3.2 Plasticity Model Calibration
The GTN matrix material is assumed to be a von Mises material. We suggest that the matrix
material follows that of Eq. (5-6). The advantages of using such yield criterion are its simplicity
and ease of calibration. In essence; the axial symmetric round bars in tension and compression
exhibit a Lode angle parameter of unity ΞΈ = 1 β 1 consecutively. On the other hand, flat grooved
specimens in tension (plane strain) exhibit a Lode angle of ΞΈ = 0.
The Lode angle dependence on material plasticity is evident for the (16 HRC) flat grooved
specimens as shown in Figure 5-6 where the dashed curves (GTN model) indicates much higher
strength than the experimental curves (solid curves). Using cs = 0.9 (i.e., including Lode angle
dependence) results in more realistic strengths (dotted curves) comparable to the experimental
tests. Furthermore, it was observed that the GTN model is quite sufficient in capturing the plasticity
behavior for the cases of 32 HRC and 39 HRC. This indicates that the material at (32 & 39 HRC)
hardness has negligible Lode angle dependence when using the GTN model, therefore cs of unity
is chosen (i.e., retrieving the original GTN model). Please refer to Table 5-3 for detail description
of plasticity parameters chosen for modified GTN. Also, please note that small compressive
Page 112
97
cylinder tests where only performed at 16 HRC due to machine capability limits. Fracture did not
occur at this hardness (16 HRC).
Table 5-3: Modified GTN plasticity parameters for 16 HRC, 32 HRC, and 39 HRC consecutively
Hardness ππ ππ ππ ππ ππ΅ πΊπ΅ ππ΅ ππ¬ π
π ππ
16 HRC 1.5 1 π12 0.0065 0.01 0.1 0.1 0.9 0.95 1
32 HRC 1.5 1 π12 0.0065 0.005 0.1 0.1 1 N/A 1
39HRC 1.5 1 π12 0.0065 0.005 0.1 0.1 1 N/A 1
5.3.3 Calibration of Failure Criteria (GTN-MMC)
It is clear from previous discussions and illustrations that the void volume fraction at failure
is not constant under different stress states. The existence of impurities in the material due to
casting and manufacturing errors could be the cause. Other causes vary from different hardness
level within the same material as well as void geometry shape changes. Therefore, a new decoupled
fracture criteria Eq. (5-24) is developed and incorporated. The calibrated parameters are shown in
Table 5-4. The set of (c1, c2) are determined from the best fit of the axisymmetric round bar
specimens in tension. While (cs) is determined from the flat grooved specimens.
Page 113
98
Table 5-4: GTN-MMC fracture criteria
Hardness 16 HRC 32 HRC 39 HRC
ππ 0.018 0.021 0.076
οΏ½οΏ½π 0.480 0.485 0.550
οΏ½οΏ½π¬ 0.310 0.320 0.300
5.3.4 Comparison between Tests and Simulation Results
The axial symmetric round bars in tension (SB, LN, SN) determine the fracture locus at ΞΈ =
1 and the plane strain specimens (LFG, MFG, SFG) determines the locus at ΞΈ = 0. The small
compressive cylinder can give fracture data at ΞΈ = β1 but the fracture was not observed. Figure
5-7 shows the 2D fracture locus of the axial symmetric tension ΞΈ = 1 and the plane strain in tension
at ΞΈ = 0, which shows the effects of stress triaxiality and Lode angle parameters. The calibrated
3D fracture loci of different tempering conditions are shown in Figure 5-8, Figure 5-9 and Figure
5-10 consecutively.
It is found that the (GTN-MMC) fracture criterion has the capability of capturing the correct
fracture locations of the round bar specimens with reasonable accuracy as seen in Figure 5-7 (solid
curves). However, the new fracture criterion still lacks the capability of capturing the correct
fracture locations for all flat grooved specimens (dotted curves). The new criterion underestimates
the fracture location occurrence for some of these specimens. However, the new decoupled (GTN-
MMC) fracture criterion is still better in capturing a more realistic fracture location for plane strain
specimens than the original GTN criteria. Please refer to Figure 5-3, Figure 5-5 and Figure 5-6
Page 114
99
where diamond, cross, and asterisk denote the GTN, (GTN-MMC) and experimental fracture
locations respectively.
It is also found that round bar specimens under tension experience larger void volume
fractions at failure when compared to plane strain (flat grooved) specimens as seen in Figure 5-7
This means that a calibrated GTN fracture parameter set (ππ , ππ) for round bar specimens will, in
general, overestimate fracture locations for plane strain specimens as previously mentioned.
Figure 5-7: Calibrated 2D fracture locus of axial symmetric tension (round bar specimens) at =
1 and plane strain tension (flat grooved specimens) at = 0
Page 115
100
Figure 5-8: Calibrated 3D fracture locus of AISI 4340 steel (16 HRC).
Figure 5-9: Calibrated 3D fracture locus of AISI 4340 steel (32 HRC).
Page 116
101
Figure 5-10: Calibrated 3D fracture locus of AISI 4340 steel (39 HRC).
Page 117
102
CHAPTER 6: STUDY OF DUCTILE FRACTURE FROM MATERIAL MICROSCOPIC IMAGES
Fractured surfaces of round bar (SB, LN, and SN) and plane strain (LFG, MFG, SFG)
specimens of hardness (16 HRC, 32 HRC, and 39 HRC) were selected for microstructural analysis
using a scanning electron microscope (SEM). Fractured surfaces of round bars reveal that the state
of stress affects the microvoids the material encounters. In essence, Figure 6-1 cases (a-c), (d-f)
and (g-h) indicate that the average microvoid size increases under an increasing stress triaxiality
state. Larger void sizes could explain the loss of ductility the material experiences under higher
stress triaxiality states. Figure 6-1 also demonstrates that the initial hardness of the material affects
void sizes. It is observed that the voids experience an increase in size for harder surface (i.e., please
refer to Figure 6-1 cases (a, d, g), (b, e) and (c, f, h)). For different magnifications of fractured
round bar specimens, please refer to Figure 6-3 through Figure 6-10.
Fractured surfaces of plane strain (flat grooved) specimen donβt reveal much size change in
the microvoids (Figure 6-2). It seems that the dominant failure mechanism for these specimens is
void shear. However, it is difficult to assess and distinguish the actual failure mechanism without
x-ray tomography testing methods. A study of in situ x-ray tomography similar to the work of
(Landron 2011) but for plane strain specimens is highly recommended for future studies. For
different magnifications of fractured flat grooved specimens, please refer to Figure 6-11 through
Figure 6-18.
Page 118
103
Figure 6-1: Fracture surfaces of round bar specimens for 16 HRC, 32 HRC, and 39 HRC consecutively
Page 119
104
Figure 6-2: Fracture surfaces of flat grooved specimens for 16 HRC, 32 HRC, and 39 HRC consecutively
Page 120
105
Figure 6-3: Fractured smooth bar specimen at different magnifications (16 HRC)
Page 121
106
Figure 6-4: Fractured smooth bar specimen at different magnifications (32 HRC)
Page 122
107
Figure 6-5: Fractured smooth bar specimen at different magnifications (39 HRC)
Page 123
108
Figure 6-6: Fractured Large Notch specimen at different magnifications (16 HRC)
Page 124
109
Figure 6-7: Fractured Large Notch specimen at different magnifications (32 HRC)
Page 125
110
Figure 6-8: Fractured sharp notch specimen at different magnifications (16 HRC)
Page 126
111
Figure 6-9: Fractured sharp notch specimen at different magnifications (32 HRC)
Page 127
112
Figure 6-10: Fractured sharp notch specimen at different magnifications (39 HRC)
Page 128
113
Figure 6-11:Fractured large flat grooved specimen at different magnifications (16 HRC)
Page 129
114
Figure 6-12: Fractured large flat grooved specimen at different magnifications (32 HRC)
Page 130
115
Figure 6-13: Fractured large flat grooved specimen at different magnifications (39 HRC)
Page 131
116
Figure 6-14: Fractured medium flat grooved specimen at different magnifications (16 HRC)
Page 132
117
Figure 6-15: Fractured medium flat grooved specimen at different magnifications (32 HRC)
Page 133
118
Figure 6-16: Fractured small flat grooved specimen at different magnifications (16 HRC)
Page 134
119
Figure 6-17: Fractured small flat grooved specimen at different magnifications (32 HRC)
Page 135
120
Figure 6-18: Fractured small flat grooved specimen at different magnifications (39 HRC)
Page 136
121
CHAPTER 7: CONCLUSION AND FUTURE WORK
7.1 Summary of Contributions
Many milestones have been achieved since the beginning of this study. These include
specimensβ design, heat-treatments, experimental tests, FEA simulations, data acquisition and data
analysis leading to results that have vastly enriched the topic. Below is a summary of the research:
A full round of numerical and experimental studies was performed to investigate the
plasticity and fracture of AISI 4340 steel alloy under multiaxial stress loading conditions
at different heat-treated conditions. The models that were used to investigate this material
were the following:
a) A modified form of the Bai-Wierzbicki (also named as PPL) model couple with a
modified Mohr-Coulomb (MMC) fracture criterion.
b) The linear Drucker-Prager Plasticity model.
c) The Gurson, Tvergaard, and Needleman (GTN) model, also known as the porous
metal plasticity model.
d) A new decoupled form of the GTN representing plasticity and fracture (GTN-
MMC) is proposed and validated.
It is found that this material exhibits both stress triaxiality and Lode angle dependent
behaviors, especially for the case of 16 Rockwell Hardness (HRC). Conversely, Lode
angle dependence for the (32 & 39 HRC) heat-treated cases was found to be less
significant. Furthermore, it was observed that the shape of the yield surface of this
Page 137
122
material changes with heat-treatment effects. The PPL model for the unheated case (16
HRC) is contained within Tresca and von Mises yield surfaces. The results of the heat-
treated case (32 HRC) exhibit a yield surface shape change and exceeds the von Mises
envelope. Further heat-treatment (39 HRC) exhibits similar effects to that of the unheated
(16 HRC). This confirms previous studies that heat-treatment not only affects mechanical
properties but also changes the yield surface shape. Thus, it is found that the heat-
treatment increases the strength while ductility is sacrificed. The 3D fracture surface of
this material is revealed, and it is found that heat-treatment had no major significance in
changing the shapes of 3D fractured surface (16, 32 and 39 HRC).
It is found that the GTN model captures the plasticity part of this material with reasonable
accuracy for round bar specimens (SB, LN, and SN). Conversely, the GTN model doesnβt
capture the plasticity part of the 16 HRC plane strain specimens (LFG, MFG, and SFG)
with reasonable accuracy. Furthermore, it was observed that the current fracture
parameter set (ππ , ππ) in GTN model donβt take into consideration specimens
experiencing different stress states, (for example calibrating (ππ , ππ) for the smooth bar
will result in underestimating fracture locations for both large notch and sharp notch also
overestimating fracture locations of plane strain specimens). Therefore, a new decoupled
fracture criterion (GTN-MMC) that is both Lode angle and stress triaxiality dependent is
developed and implemented. This is done by incorporating the Bai and Wierzbicki yield
criterion, which incorporates Lode angle dependence, as the matrix material instead of
the traditional J2 plasticity used in the GTN. The (GTN-MMC) fracture criterion has the
Page 138
123
capability of capturing the correct fracture locations of round bar specimens with
reasonable accuracy. However, the new fracture criterion still lacks the capability of
capturing the correct fracture locations for all flat grooved specimens respectively. The
new criterion underestimates the fracture location occurrence. However, the new
decoupled (GTN-MMC) fracture criterion is much better in capturing a more realistic
fracture location for plane strain specimens than the original GTN criteria.
Fracture surface analysis of round bars using (SEM) reveals that the state of stress affects
the microvoids the material encounters, i.e., larger microvoids appear for higher stress
triaxiality encountered. While, fractured surfaces of plane strain (flat grooved) specimen
donβt reveal much size change in the microvoids, which is probably due to microvoid
shearing effect.
Page 139
124
7.2 Recommended Future Study
Ductile fracture after large plastic deformation become a more and more important subject
as the requirements of high fidelity modeling in computers. The microvoid based fundamental
mechanisms are not fully understood and they are also critical to future material design and
development. Below is a list of suggested topics for future study.
1. Effects of strain rate, temperature and anisotropy: The PPL-MMC plasticity/fracture
model describes the materials behaviors under different loading conditions considering
various stress triaxiality and Lode angle effects. This model has also the capability of
correctly capturing various fracture propagation phenomena which makes it a unique
candidate for such a study. However, the coupling effects of strain rate and temperature
have not fully addressed. There is also a need to include material anisotropic effect.
2. Heat-treatment effects on Lode angle parameter: Heat-treatment on AISI 4340 steel
seems to restrict the Lode angle dependence behaviors. Heat-treatment and Lode angle
dependence should be investigated using different materials.
3. X-ray tomography: To better understand the microvoid behaviors and attributes under
different loading conditions, A study similar to the work of (Landron 2011) for different
loading conditions including axisymmetric smooth and notched round bars, plane strain,
and shear specimens are highly recommended for future study.
4. Void behaviors and attributes: Void behaviors are complex and different under different
stress states. It has been recognized that the void nucleation, growth, and coalescence are
the primary attributes at high-stress triaxiality. However, it is also known that the void
Page 140
125
shape at low or even negative stress triaxiality is entirely different. The issue of void shape
within the unit cell modeling has been gaining much attention recently. The current GTN
model does not account for this. (Wen, Hwang et al. 2005) extended the GTN model to
account for such dependency. However, the resulting equations are complicated and hard
to implement for practical cases. An applicable and easy way to implement criterion should
be addressed.
Page 141
126
Here are the papers that I have published or submitted for publication:
1. Jia Y, Ghazali S, Bai Y. Application of eMMC Model to Fracture of Metal Sheets. In Fracture,
Fatigue, Failure and Damage Evolution, Volume 8 2017 (pp. 49-55). Springer International
Publishing.
2. Ghazali, S., Choi, Y., Bai, Y., βStudy on the plasticity and fracture of AISI 4340 steel under
different loading conditions and considering heat-treatment effectsβ, International Journal of
Fracture, submitted for publication.
3. Ghazali S, Bai Y. Application of Modified Gurson-Tvergaard Needleman (GTN) to AISI 4340
steel, Engineering Fracture Mechanics, (manuscript ready).
Page 142
127
:
FLOW RULE DERIVATION CONSIDERING ππ(οΏ½οΏ½)
Page 143
128
The plastic flow rule of the Bai-Wierzbicki (also called PPL) plasticity model reads as follows:
where π , οΏ½οΏ½(νπ), πΞ·, οΏ½οΏ½, π0, πππ , π, πΎ and ππ
ππ₯ are parameters defined in section 4.2
Deriving the plastic flow rule (π) with respect to πππ the following is obtained:
ππ
ππππ=ππ
ππππ+ οΏ½οΏ½(νπ) [
πππ
ππππ (οΏ½οΏ½ β π0) + ππ
ποΏ½οΏ½
ππππ] [ππ
π
+ (ππππ₯ β ππ
π ) (πΎ βπΎπ+1
π + 1)(1 +
1
π)
β οΏ½οΏ½(νπ) [1 β ππ(οΏ½οΏ½ β π0)]] (ππππ₯ β ππ
π ) (1 +1
π) (1 β πΎπ)
ππΎ
ππππ
(A-2)
Assuming that the πΞ· is a function of Lode angle πΞ· is expressed as:
πΞ· = π1 β π2 π(οΏ½οΏ½) (A-3)
Where, π1, π2 are constant parameters.
π(οΏ½οΏ½) =1
1 + πβπ( π0βοΏ½οΏ½)
(A-4)
Deriving πΞ· with respect to πππ the following is obtained:
π πΞ·
ππππ= β π2
ππ(οΏ½οΏ½)
ππππ
(A-5)
ππ(οΏ½οΏ½)
ππππ is expressed as follows:
π = π β οΏ½οΏ½(νπ) [1 β πΞ· (οΏ½οΏ½ β π0)] [πππ + (ππ
ππ₯ β πππ )(πΎ β
πΎπ+1
π+ 1 ) (1 +
1
π)]
(A-1)
Page 144
129
ππ(οΏ½οΏ½)
ππππ=
βπ ποΏ½οΏ½ππππ
πβπ( π0βοΏ½οΏ½)
(πβπ( π0βοΏ½οΏ½) + 1)2
(A-6)
Substituting Eq. (A-6) into Eq. (A-5):
π πΞ·
ππππ= β π2
βπ ποΏ½οΏ½ππππ
πβπ( π0βοΏ½οΏ½)
(πβπ( π0βοΏ½οΏ½) + 1)2
(A-7)
The Lode angle parameter οΏ½οΏ½ is expressed as the following:
οΏ½οΏ½ = 1 β2
π cosβ1 π
(A-8)
Deriving οΏ½οΏ½ with respect to πππ the following is obtained:
ποΏ½οΏ½
ππππ=
2
πβ1 β π2 ππ
ππππ
(A-9)
Substituting Eq. (A-9) into Eq. (A-7) :
π πΞ·
ππππ=
2 π2 π πβπ( π0βοΏ½οΏ½)
πβ1 β π2 (πβπ( π0βοΏ½οΏ½) + 1)2 ππ
ππππ
(A-10)
π is defined as follows:
π = (π
π)3
(A-11)
Deriving π with respect to πππ the following is obtained:
ππ
ππππ=
3 π2 (π ππππππ
β π ππππππ
)
π4
(A-12)
The goal here is to find ππ
ππππ and
ππ
ππππ therefore π and π½3 are defined as follows:
π = (9
2 π . [π] βΆ [π] )
13β
(A-13)
Page 145
130
π½3 =1
3 πππ πππ πππ = det [π]
(A-14)
Expressing π as a function of π½3 the following is obtained:
π = (13.5 π½3 )13β = (13.5 det [π] )
13β (A-15)
Now we need to express π as a function of π½2 and it is expressed as follows:
π½2 = 1
2 πππ πππ =
π2
3
(A-16)
Taking derivatives of the previous parameters results in the following:
ππ½3ππππ
= πππ πππ β1
3 πππ πππ πΏππ = πππ πππ β
2
3 π½2 πΏππ
(A-17)
ππ
ππππ= (13.5 )
13β( π½3 )
3
β23β ππ½3ππππ
= (13.5 )13β ( π½3 )
3
β23β
(πππ πππ β2
3 π½2 πΏππ)
(A-18)
ππ
ππππ= (
1
2 )
13β
( det [π] )β2
3β ( πππ πππ β2
3 π½2 πΏππ )
(A-19)
ππ
ππππ=
3
2 π π ππ
(A-20)
From Eqn. (A-2) the ποΏ½οΏ½
ππππ and
ππΎ
ππππ parameters still need to be found.
They are expressed as follows:
ποΏ½οΏ½
ππππ=πΏππ3 πβ
(A-21)
ππΎ
ππππ = (
3β3
2 β β3)(
tan(π βπ6)
cos(π βπ6))
1
π sin (3π) (πΏππ3+cos (3π)
2π π ππ β
3
2π2π πππ ππ)
(A-22)
Page 146
131
Please note that to satisfy the assumption of plastic incompressibility, the term πΏππ
3 πβ in Eqn. (A-21)
is removed.
Substituting Eqns. (A-19) and (A-20) into Eq. (A-12):
ππ
ππππ=
3 π2(π((12)
13( det[π] )
β23 (πππ πππ β
29 π
2 πΏππ )) β π (3π ππ2 π ))
π4
(A-23)
Substituting Eqns.(A-15) into Eq. (A-23):
ππ
ππππ
=
(13.5 det[π])23(π (
12)
13(det[π])
β23 (ππππππ β
2π2πΏππ9
) β (13.5 det[π])13 (3π ππ2 π ))
π4
3
(A-24)
Substituting Eq. (A-24) into (A-10)
π πΞ·
ππππ= (
2π2 π πβπ( π0βοΏ½οΏ½)
πβ1 β π2 (πβπ( π0βοΏ½οΏ½) + 1)2)
Γ
(
(13.5 det[π])
23(π (
12)
13(det[π])
β23 (ππππππ β
2π2πΏππ9
) β (13.5 det[π])13 (3π ππ2 π ))
π4
3
)
(A-25
)
Substituting Eqns. (A-20), (A-21) and (A-22) into (A-2) the following is obtained:
Page 147
132
ππ
ππππ=3π ππ2π
+ οΏ½οΏ½(νπ)
[ 27π2 π π
βπ( π0βοΏ½οΏ½) π2 (ππππππ β2π2πΏππ9
) β 60.75 π ππ det[π]
πβ1 β π2 (πβπ( π0βοΏ½οΏ½) + 1)2π5
(οΏ½οΏ½ β π0)
]
[πππ
+ (ππππ₯ β ππ
π ) (πΎ βπΎπ+1
π + 1)(1 +
1
π) β οΏ½οΏ½(νπ)[1 β ππ(οΏ½οΏ½ β π0)]] (ππ
ππ₯ β πππ ) (1
+1
π) (
3β3 (1 β πΎπ)
π sin (3π)(2 β β3))(
tan(π βπ6)
cos(π βπ6))(
πΏππ3+cos(3π) π ππ
2π β3 π πππ ππ2π2
)
(A-26)
To satisfy the assumption of plastic incompressibility, the term πΏππ
3 is also removed. Therefore, a deviatoric associative flow rule
is developed as follows:
πνππ = ππ [ππ
ππππ]
(A-27)
Page 148
133
πνππ = ππ
[ 3π ππ
2π
+ οΏ½οΏ½(νπ)
[ 27π2 π π
βπ( π0βοΏ½οΏ½) π2 (ππππππ β2π2πΏππ9
) β 60.75 π ππ det[π]
πβ1 β π2 (πβπ( π0βοΏ½οΏ½) + 1)2π5
(οΏ½οΏ½ β π0)
]
[πππ
+ (ππππ₯ β ππ
π ) (πΎ βπΎπ+1
π + 1)(1 +
1
π) β οΏ½οΏ½(νπ)[1 β ππ(οΏ½οΏ½ β π0)]] (ππ
ππ₯ β πππ ) (1
+1
π) (
3β3 (1 β πΎπ)
π sin (3π)(2 β β3))(
tan(π βπ6)
cos(π βπ6))(
cos(3π) π ππ2π
β3 π πππ ππ2π2
)
]
(A-28)
The present formulation can be classified into an intermediate flow rule between the fully associated and non-associated flow
rule.
Page 149
134
: DETERMINING TRUE STRESS-STRAIN CURVES BASED ON BRIDGMANβS ROUND
NOTCHED BARS USING DIGITAL IMAGING CORRELATION (DIC)
Page 150
135
The Goal of this section is to determine the evolution of the minimum cross section βπΌβ (p2p5)
and the change in radii (O1p2 & O2p5) shown in Figure B-1. Knowing this allows the
determination of true stress at every time interval including after necking by using Bridgman
solutions for round notched specimens.
Figure B-1: The cross-section of a notched specimen
Page 151
136
The following points are defined as follows:
Point/vector name Coordinate system correspondence
O1 (x0,y0)
O2 (xf,yf)
p1 (x1,y1)
p2 (x2,y2)
p3 (x3,y3)
p4 (x4,y4)
p5 (x5,y5)
p6 (x6,y6)
O1p1 (π₯1 β π₯0 , π¦1 β π¦0)
O1q1 (π₯1 + π₯22
β π₯0 ,π¦1 + π¦22
β π¦0)
O1p2 (Radius 1) (π₯2 β π₯0 , π¦2 β π¦0)
O1q2 (π₯3 + π₯22
β π₯0 ,π¦3 + π¦22
β π¦0)
O1p3 (π₯3 β π₯0 , π¦3 β π¦0)
O2p4 (π₯4 β π₯π , π¦4 β π¦π)
O2q3 (π₯4 + π₯52
β π₯π ,π¦4 + π¦52
β π¦π)
O2p5 (Radius 2) (π₯5 β π₯π , π¦5 β π¦π)
O2q4 (π₯5 + π₯62
β π₯π ,π¦5 + π¦62
β π¦π)
O2p6 (π₯6 β π₯π , π¦6 β π¦π)
Page 152
137
Since π1π1 β₯ q1p2 πππ π1π2 β₯ q2p3
β΄ (O1q1 ) . ( q1p2 ) = 0 and (O1q2 ) . ( q2p3 ) = 0
The solution of this equation with respect to (x0, y0) defines the change of (x0, y0) with respect
to the other parameters which are known at every time step. (x0, y0) are read as follows:
π₯0 =x12 β x2
2 + y12 β y2
2 +2 y1a1a2
β 2 y2 a1a2
2 x1 β 2 x2
(B-1)
where,
π1 = (π¦22 β
π¦32) (
π¦22+ π¦32) + (
π₯22 β
π₯32) (π₯22+ π₯32βπ₯12 β π₯2
2 + π¦12 β π¦2
2
2 π₯1 β 2 π₯2)
(B-2)
π2 = (π¦32β π¦22) +
(π₯22 β
π₯32) (2 π¦1 β 2 π¦2)
2 π₯1 β 2 π₯2
(B-3)
π¦0 =(π¦22 β
π¦32) (
π¦22+ π¦32)+π4 (
π₯22+ π₯32βπ₯12 β π₯2
2 + π¦12 β π¦2
2
π3)
(π¦32β π¦22) +
π4(2 π¦1 β 2 π¦2)π3
(B-4)
where,
π3 = 2 π₯1 β 2 π₯2 (B-5)
Page 153
138
π4 = (π₯22 β
π₯32)
(B-6)
Radius1 reads as:
π
1 = β(π₯1 βπ₯12 β π₯2
2 + π¦12 β π¦2
2 +2 π¦1(π5)π6
β 2 π¦2(π5)π6
2 π₯1 β 2 π₯2)
2
+ (π¦1 +π5π6)2
(B-7)
where,
π5 = (π¦22 β
π¦32) (
π¦22+ π¦32) + (
π₯22 β
π₯32) (π₯22+ π₯32βπ₯12 β π₯2
2 + π¦12 β π¦2
2
2 π₯1 β 2 π₯2)
(B-8)
π6 = (π¦32β π¦22) +
(π₯22 β
π₯32) (2 π¦1 β 2 π¦2)
2 π₯1 β 2 π₯2
(B-9)
The same argument goes for the left side of the notch (xf, yf) and radius 2 read as follows:
π₯π =π₯4
2 β π₯52 + π¦4
2 β π¦52 +
2 π¦4(π1)π2
β 2 π¦5(π1)π2
2 π₯4 β 2 π₯5
(B-10)
where,
Page 154
139
π1 = (π¦52 β
π¦62) (
π¦52+ π¦62) + (
π₯52 β
π₯62) (π₯52+ π₯62βπ₯4
2 β π₯52 + π¦4
2 β π¦52
2 π₯4 β 2 π₯5)
(B-11)
π2 = (π¦62β π¦52) +
(π₯52 β
π₯62) (2 π¦4 β 2 π¦5)
2 π₯4 β 2 π₯5
(B-12)
π¦π =(π¦52 β
π¦62) (
π¦52+ π¦62)+π4 (
π₯52+ π₯62βπ₯4
2 β π₯52 + π¦4
2 β π¦52
π3)
(π¦62β π¦52) +
π4(2 π¦4 β 2 π¦5)π3
(B-13)
where,
π3 = 2 π₯4 β 2 π₯5 (B-14)
π4 = (π₯52 β
π₯62)
(B-15)
Radius 2 reads as:
R2 = β(x4 βx42 β x52 + y42 β y52 +
2 y4(b5)a6
β 2 y5(b5)b6
2 x4 β 2 x5)
2
+ (y4 +b5b6)2
(B-16)
Page 155
140
where,
π5 = (π¦52 β
π¦62) (
π¦52+ π¦62) + (
π₯52 β
π₯62) (π₯52+ π₯62βπ₯4
2 β π₯52 + π¦4
2 β π¦52
2 π₯4 β 2 π₯5)
(B-17)
π6 = (π¦62β π¦52) +
(π₯52 β
π₯62) (2 π¦4 β 2 π¦5)
2 π₯4 β 2 π₯5
(B-18)
Since both radii are known as a function of 6 known points the minimum cross section βπΌβ is
known and reads as follows:
πΌ =1
2β(|π1 β π2|
2 β |π5π6βπ5π6|2
) β1
2β(|π₯4 β π1|
2 β |π¦4 +π5π6|2
)
β 1
2β(|π₯1 β π2|
2 β |π¦1 +π5π6|2
)
(B-19)
where,
π1 = (π₯4
2 β π₯52 + π¦4
2 β π¦52 +
2 π¦4(π5)π6
β 2 π¦5(π5)π6
2 π₯4 β 2 π₯5)
(B-20)
Page 156
141
π2 = (π₯12 β π₯2
2 + π¦12 β π¦2
2 +2 π¦1(π5)π6
β 2 π¦2(π5)π6
2 π₯1 β 2 π₯2)
(B-21)
The changes in radii and minimum cross-sectional thicknesses are known at every time step. The
user must track six points in the DIC. True stress strain is readily determined by Bridgman stress-
strain solutions.
Page 157
142
Matlab Code:
%Clearing commands
clc
format compact
clear all
syms x1 x2 x3 x0 y0 y1 y2 y3 xa ya x4 y4 x5 y5 x6 y6 xf yf
% --------------------------------------------------------------
% dealing with the three points on the right plane (p1,p2,p3)
oq1=[(x1+x2)/2-x0,(y2+y1)/2-y0];
q1p2=[x2-(x1+x2)/2,y2-(y1+y2)/2];
dot1x=oq1(1,1)*q1p2(1,1);
dot1y=oq1(1,2)*q1p2(1,2);
dot1=dot1x+dot1y;
new_x0=solve(dot1==0,x0);
oq2=[(x2+x3)/2-x0,(y2+y3)/2-y0];
q2p3=[x3-(x3+x2)/2,y3-(y3+y2)/2];
dot2x=oq2(1,1)*q2p3(1,1);
dot2y=oq2(1,2)*q2p3(1,2);
dot2=dot2x+dot2y;
Page 158
143
% now solving for x0,y0 in terms of three points selected
dot2_new=subs(dot2,x0,new_x0);
new_y0=solve(dot2_new==0,y0);
las_x0=subs(new_x0,y0,new_y0);
% now solving for radius in terms of selected points
radius_x1=x1-las_x0;
radius_y1=y1-new_y0;
radius=[radius_x1,radius_y1];
radius=norm(radius);
% --------------------------------------------------------------
%now dealing with the other 3points on the left plane (p4,p5,p6)
%same argument performed on the right plane points is done here
oq3=[(x4+x5)/2-xf,(y5+y4)/2-yf];
q3p5=[x5-(x4+x5)/2,y5-(y4+y5)/2];
dot3x=oq3(1,1)*q3p5(1,1);
dot3y=oq3(1,2)*q3p5(1,2);
dot3=dot3x+dot3y;
new_xf=solve(dot3==0,xf);
oq4=[(x5+x6)/2-xf,(y5+y6)/2-yf];
q4p6=[x6-(x6+x5)/2,y6-(y6+y5)/2];
dot4x=oq4(1,1)*q4p6(1,1);
Page 159
144
dot4y=oq4(1,2)*q4p6(1,2);
dot4=dot4x+dot4y;
dot4_new=subs(dot4,xf,new_xf);
new_yf=solve(dot4_new==0,yf);
las_xf=subs(new_xf,yf,new_yf);
radius_2_x=x4-las_xf;
radius_2_y=y4-new_yf;
radius_2=[radius_2_x,radius_2_y];
radius_2=norm(radius_2);
%now determining the total length from point O1 to O2
length_tot=[x0-xf,y0-yf];
length_tot=norm(length_tot);
%now determining the minimum cross-sectional thickness
a=length_tot-radius_2-radius;
a=subs(a,xf,las_xf);
a=subs(a,yf,new_yf);
a=subs(a,x0,las_x0);
a=subs(a,y0,new_y0);
% a is redefined as half thickness
a=0.5*a;
%%%%%%%%%%%%%%%%%%%%%%
Page 160
145
fprintf(' x0 =\n\n')
pretty(las_x0)
fprintf(' y0 =\n\n')
pretty(new_y0)
fprintf(' radius_1 =\n\n')
pretty(radius)
fprintf(' xf =\n\n')
pretty(las_xf)
fprintf(' yf =\n\n')
pretty(new_yf)
fprintf(' radius_2 =\n\n')
pretty(radius_2)
fprintf(' a =\n\n')
pretty(a)
% --------------------------------------------------------------
% material parameter input
load('force_sb.mat')
load('data_sb.mat')
% specifying all 6 points (x, y) coordinates
while (i < 14)
i=i+1;
Page 161
146
x1=data_sb(i,1);
y1=data_sb(i,2);
x2=data_sb(i,3);
y2=data_sb(i,4);
x3=data_sb(i,5);
y3=data_sb(i,6);
x4=data_sb(i,7);
y4=data_sb(i,8);
x5=data_sb(i,9);
y5=data_sb(i,10);
x6=data_sb(i,11);
y6=data_sb(i,12);
gage_length_start=data_sb(1,4)-data_sb(1,2);
gage_length_step(i)=data_sb(i,4)-data_sb(i,2);
delta(i)=gage_length_step(i)-gage_length_start;
a_evolv(i)=abs(eval(a)).*10^-3; % in m
radius_evolv(i)=eval(radius).*10^-3; % in m
radius_evolv_2(i)=eval(radius_2).*10^-3; % in m
% to determine true stress-strain at every point
true_stress=0; %initially
a_initial=0.00254;
Page 162
147
if true_stress <=838.48
eng_strain(i)=delta(i)./gage_length_start;
true_strain(i)=log(1+eng_strain(i));
eng_stress(i)=force_sb(1,i)./(pi.*a_initial.^2);
true_stress(i)=eng_stress(i).*(1+eng_strain(i));
true_plastic_strain(i)=true_strain(i)-true_stress(i)./204*10^-3;
else
num(i)=force_sb(1,i)./(pi*a_evolv(i).^2);
den(i)=(1+2*radius_evolv(i)/a_evolv(i))*log(1+a_evolv(i)/(2*...
radius_evolv(i)));
%if radius_evolv(i)>=0 & radius_evolv_2(i)>=0
stress_bridgman(i)=num(i)/den(i)
plastic_strain(i)=2*log(abs(a_evolv(i))./0.00254)*-1;
% end
scatter(plastic_strain,stress_bridgman)
end
end
Page 163
148
: PLASTIC FLOW POTENTIAL DERIVATION FOR PROPOSED MODIFIED GTN
MODEL
Page 164
149
The plastic flow rule of the Gurson-Tvergaard-Needleman (GTN) model reads as follows:
F = q β π¦β(1 + π3πβ2)β 2π1π
β cosh (β3π2π
2π¦) = 0
(C-1)
The original GTN model assumes π¦ to follow the Von Mises criterion. Here we assume that the
π¦ is expressed as follows:
π¦ = π¦ [cs + (c
ax β cs) (
π + 1
π)(β
m+1
m+ 1)]
(C-2)
where π , πππ , π, , ππ
ππ₯ are parameters defined in section 4.2.
The flow rule is reformulated into the following:
F = q β π¦ [cs + (c
ax β cs) (
π + 1
π)
β (βm+1
m+ 1)]β(1 + π3π
β2)β 2π1πβ cosh (
β3π2π
2π¦) = 0
(C-3)
Deriving F with respect to ππ we obtain the following:
πF
πππ=
ππ
πππβ
3π1 π2 πβ sinh (
3π2π 20
) πΏππ
2 β(1 + (π1πβ)2) β 2π1π
β cosh (3π2π 20
)
β [cs + (c
ax β cs) (
π + 1
π)(β
m+1
m+ 1)] +
π
πππβ π¦
β [(cax β c
s) (π + 1
π) (1 β m)]
β β(1 + π3πβ2)β 2π1π
β cosh (β3π2π
2π¦)
(C-4)
Page 165
150
The terms ππ
πππ and
π
πππ can be expressed as follows:
ππ
πππ=
3
2 qsππ
(C-5)
π
πππ= (
3β3
2 β β3)(
tan (π βπ6)
cos (π βπ6))
1
π sin (3π) (πΏππ3+cos(3π)
2π sππ β
3
2π2 sππsππ )
(C-6)
Substituting Eqn.βs (C-5) and (C-6) into (C-4) and assuming an associative flow rule the following plastic strain is developed:
πΞ΅ππππ = ππ
(
3sππ
2 q β
3π1 π2 πβ sinh (
3π2π 20
) πΏππ
2 β(1 + (π1πβ)2) β 2π1π
β cosh (3π2π 20
)
[cs + (c
ax β cs) (
π + 1
π)(β
m+1
m+ 1)]
+ ((3β3
2 β β3)(
tan (π βπ6)
cos (π βπ6))
1
π sin(3π) (πΏππ3+cos(3π)
2π sππ β
3
2π2 sππsππ )π¦
Γ [(cax β c
s) (π + 1
π) (1 β m)]β(1 + π3π
β2)β 2π1πβ cosh (
β3π2π
2π¦) )
)
(C-7)
Page 166
151
: HEAT-TREATMENT CERTIFICATIONS
Page 169
154
REFERENCES
Abbassi, F., et al. (2010). Effect of ductile damage evolution in sheet metal forming: experimental
and numerical investigations. Key Engineering Materials, Trans Tech Publ.
Afshan, S. (2013). "Micromechanical Modelling of Damage Healing in Free Cutting Steel."
Algarni, M. (2017). "A Study On The Plasticity And Fracture Behaviors Of Inconel 718 Under
Multiaxial Stress And Extremely Low Cycle Fatigue Loadings."
Algarni, M., et al. (2015). "A study of Inconel 718 dependency on stress triaxiality and Lode angle
in plastic deformation and ductile fracture." Engineering Fracture Mechanics 147: 140-157.
Algarni, M., Bai Y, Choi Y (2015). "A study of Inconel 718 dependency on stress triaxiality and
Lode angle in plastic deformation and ductile fracture." Engineering Fracture Mechanics 147: 140-
157.
Algarni, M., et al. (2017). "A unified material model for multiaxial ductile fracture and extremely
low cycle fatigue of Inconel 718." International Journal of Fatigue 96: 162-177.
Algarni, M. S., et al. (2015). "Linkage between Ductile Fracture and Extremely Low Cycle Fatigue
of Inconel 718 Under Multiaxial Loading Conditions." TMS2015 Supplemental Proceedings:
1023-1030.
Areias, P. and T. Rabczuk (2013). "Finite strain fracture of plates and shells with configurational
forces and edge rotations." International Journal for Numerical Methods in Engineering 94(12):
1099-1122.
Argon, A. and J. Im (1975). "Separation of second phase particles in spheroidized 1045 steel, Cu-
0.6 pct Cr alloy, and maraging steel in plastic straining." Metallurgical Transactions A 6(4): 839.
ASTM-STEEL. "4340 Steel ". Retrieved 23 June, 2016, from
http://www.astmsteel.com/product/4340-steel-aisi/.
Babout, L., et al. (2004). "On the competition between particle fracture and particle decohesion in
metal matrix composites." Acta Materialia 52(15): 4517-4525.
Bai, Y. (Massachusetts Institute of Technology ;2008). "Effect of loading history on necking and
fracture [PhD thesis]."
Bai, Y., Wierzbicki T (2008). "A new model of metal plasticity and fracture with pressure and
Lode dependence." International Journal of Plasticity 24(6): 1071-1096.
Page 170
155
Bai, Y. and T. Wierzbicki (2008). "A new model of metal plasticity and fracture with pressure and
Lode dependence." International Journal of Plasticity 24(6): 1071-1096.
Bai, Y. and T. Wierzbicki (2009). "Application of extended MohrβCoulomb criterion to ductile
fracture." International Journal of Fracture 161(1): 1-20.
Bai, Y., Wierzbicki T (2015). "A comparative study of three groups of ductile fracture loci in the
3D space." Engineering Fracture Mechanics 135: 147-167.
Bai, Y. and T. Wierzbicki (2015). "A comparative study of three groups of ductile fracture loci in
the 3D space." Engineering Fracture Mechanics 135: 147-167.
Bao, Y. and T. Wierzbicki (2004). "On fracture locus in the equivalent strain and stress triaxiality
space." International Journal of Mechanical Sciences 46(1): 81-98.
Becker, R., et al. (1988). "Void growth and failure in notched bars." Journal of the Mechanics and
Physics of Solids 36(3): 317-351.
Beese, A. M., et al. (2010). "Partially coupled anisotropic fracture model for aluminum sheets."
Engineering Fracture Mechanics 77(7): 1128-1152.
Bell, T. and N. Loh (1982). "The fatigue characteristics of plasma nitrided three pct Cr-Mo steel."
Journal of Heat Treating: 232-237.
Benzerga, A. A. (2000). Rupture ductile des tΓ΄les anisotropes. Simulation de la propagation
longitudinale dans un tube pressurisΓ©, Ecole Nationale SupΓ©rieure des Mines de Paris.
Benzerga, A. A., et al. (2016). "Ductile failure modeling." International Journal of Fracture 201(1):
29-80.
Beremin, F. (1981). "Cavity formation from inclusions in ductile fracture of A508 steel."
Metallurgical and Materials Transactions A 12(5): 723-731.
BΓΈrvik, T., et al. (2001). "A computational model of viscoplasticity and ductile damage for impact
and penetration." European Journal of Mechanics-A/Solids 20(5): 685-712.
Bouaziz, O., et al. (2008). Revue de MΓ©tallurgieβInternational Journal of Metallurgy 105(2): 102-
107.
Brown, L. and J. Embury (1973). The microstructure and design of alloys. Proc. 3rd Int. Conf. on
the Strength of Metals and Alloys, The Institute of Metals, London.
Page 171
156
Brown, M. and D. Embury (1973). A model of ductile fracture in two-phase materials. Proceedings
of 3rd international conference on strength of metals and alloys.
Cao, S., et al. (2014). "A Lode-dependent enhanced Lemaitre model for ductile fracture prediction
at low stress triaxiality." Engineering Fracture Mechanics 124-125: 80-96.
CastaΓ±eda, P. P. and M. Zaidman (1994). "Constitutive models for porous materials with evolving
microstructure." Journal of the Mechanics and Physics of Solids 42(9): 1459-1497.
Cazacu, O. and F. Barlat (2004). "A criterion for description of anisotropy and yield differential
effects in pressureinsensitive metals.". International Journal of Plasticity 20 (11): 2027β2045.
Chu, C. and A. Needleman (1980). "Void nucleation effects in biaxially stretched sheets." Journal
of Engineering Materials and Technology(Transactions of the ASME) 102(3): 249-256.
Chung, K., et al. (2011). "A modified damage model for advanced high strength steel sheets."
International Journal of Plasticity 27(10): 1485-1511.
Cockcroft, M. and D. Latham (1968). "Ductility and the workability of metals." J Inst Metals 96(1):
33-39.
Cox, T. and J. R. Low (1974). "An investigation of the plastic fracture of AISI 4340 and 18 Nickel-
200 grade maraging steels." Metallurgical and Materials Transactions B 5(6): 1457-1470.
Drucker, D. C. and W. Prager (1952). "Soil mechanics and plastic analysis or limit design."
Quarterly of applied mathematics 10(2): 157-165.
Dunand, M. and D. Mohr (2011). "On the predictive capabilities of the shear modified Gurson and
the modified MohrβCoulomb fracture models over a wide range of stress triaxialities and Lode
angles." Journal of the Mechanics and Physics of Solids 59(7): 1374-1394.
Erice, B. and F. GΓ‘lvez (2014). "A coupled elastoplastic-damage constitutive model with Lode
angle dependent failure criterion." International Journal of Solids and Structures 51(1): 93-110.
Fansi, J. (2013). Ecole nationale supérieure d'arts et métiers-ENSAM; Université de Liège.
Fleck, N. (1997). Strain gradient plasticity. Adv. Appt. Mech. 33 (Hutchinson, JW, Wu, TY, eds.),
New York: Academic Press.
Page 172
157
Fond, C., et al. (1996). "Polymers toughened with rubber microspheres: an analytical solution for
stresses and strains in the rubber particles at equilibrium and rupture." International Journal of
Fracture 77(2): 141-159.
Fossum, F. and M. Brannon (2006). "On a viscoplastic model for rocks with mechanism-dependent
characteristic times." Acta Geotechnica 1(2): 89-106.
Foster, T., et al. (2011). "Dynamic crack initiation toughness of 4340 steel at constant loading
rates." Engineering Fracture Mechanics 78(6): 1264-1276.
Gachet, M., et al. (2014). "Fracture mechanisms under monotonic and non-monotonic low Lode
angle loading." Engineering Fracture Mechanics 124-125: 121-141.
Gao, X., et al. (1998). "Ductile tearing in part-through cracks: experiments and cell-model
predictions." Engineering Fracture Mechanics 59(6): 761-777.
Gao, X. and J. Kim (2006). "Modeling of ductile fracture: significance of void coalescence."
International Journal of Solids and Structures 43: 6277β6293.
Gao, X., et al. (2005). "On ductile fracture initiation toughness: Effects of void volume fraction,
void shape and void distribution." International Journal of Solids and Structures 42(18-19): 5097-
5117.
Gao, X., Zhang T, Hayden M, Roe C (2009). "Effects of the stress state on plasticity and ductile
failure of an aluminum 5083 alloy." International Journal of Plasticity 25(12): 2366-2382.
Garrison, W. and N. Moody (1987). "Ductile fracture." Journal of Physics and Chemistry of Solids
48(11): 1035-1074.
Ghajar, R., et al. (2012). "Sensitivity Analysis on Triaxiality Factor and Lode Angle in Ductile
Fracture." Journal of Mechanics 29(01): 177-184.
Gologanu, M., et al. (1993). "Approximate models for ductile metals containing non-spherical
voidsβcase of axisymmetric prolate ellipsoidal cavities." Journal of the Mechanics and Physics
of Solids 41(11): 1723-1754.
Gologanu, M., et al. (1995). "Recent extensions of Gursonβs models for ductile porous metals."
Continuum mechanics. Berlin: Springer.
Goods, S. and L. Brown (1979). "Overview No. 1: The nucleation of cavities by plastic
deformation." Acta metallurgica 27(1): 1-15.
Page 173
158
Gruben, G., et al. (2012). "Evaluation of uncoupled ductile fracture criteria for the dual-phase steel
Docol 600DL." International Journal of Mechanical Sciences 62(1): 133-146.
Gurson, A. L. (1975). Plastic flow and fracture behavior of ductile materials, incorporating void
nucleation, growth, and interaction, Brown University Providence, RI.
Gurson, A. L. (1977). Journal of engineering materials and technology 99(1): 2-15.
Gurson, A. L. (1977). "Continuum theory of ductile rupture by void nucleation and growth: Part
IβYield criteria and flow rules for porous ductile media." Journal of Engineering Materials and
Technology 99(1): 2-15.
HΓ
KAN, H. (2011). "Warhead penetration in concrete protective structures." thesis.
Hancock, J. and D. Brown (1983). "On the role of strain and stress state in ductile failure." Journal
of the Mechanics and Physics of Solids 31(1): 1-24.
Hancock, J. and A. Mackenzie (1976). "On the mechanisms of ductile failure in high-strength
steels subjected to multi-axial stress-states." Journal of the Mechanics and Physics of Solids 24(2-
3): 147-160.
Hancock, W. and C. Mackenzie (1976). "On the mechanisms of ductile failure in high-strength
steels subjected to multi-axial stress-states." Journal of the Mechanics and Physics of Solids 24(2-
3): 147-160.
Hansson, H. (2011). Warhead penetration in concrete protective structures, KTH Royal Institute
of Technology.
Hickey, F. and A. Anctil (1985). "Split heat mechanical property comparison of ESR and VAR
4340 steel." Journal of Heat Treating 4(2): 177-183.
Hill, R. (1984). On macroscopic effects of heterogeneity in elastoplastic media at finite strain.
Mathematical Proceedings of the Cambridge Philosophical Society, Cambridge University Press.
Huo, B., et al. (1999). "A note on the effect of surface energy and void size to void growth."
European Journal of Mechanics-A/Solids 18(6): 987-994.
Jia, Y. and Y. Bai (2016). "Ductile fracture prediction for metal sheets using all-strain-based
anisotropic eMMC model." International Journal of Mechanical Sciences 115: 516-531.
Jia, Y., et al. (2017). Application of eMMC Model to Fracture of Metal Sheets. Fracture, Fatigue,
Failure and Damage Evolution, Volume 8, Springer: 49-55.
Page 174
159
Jia, Y., et al. (2017). "A Comprehensive Plasticity and Fracture Model for Metal Sheets under
Multi-axial Stress and Non-Linear Strain Path." SAE International Journal of Engines 10(2017-
01-0315): 266-273.
Johnson, G. R. and W. H. Cook (1985). "Fracture characteristics of three metals subjected to
various strains, strain rates, temperatures and pressures." Engineering Fracture Mechanics 21(1):
31-48.
Kanetake, N., et al. (1995). "Continuous observation of microstructural degradation during tensile
loading of particle reinforced aluminium matrix composites." Materials science and technology
11(12): 1246-1252.
Kiran, R. and K. Khandelwal (2014). Fatigue & Fracture of Engineering Materials & Structures
37(2): 171-183.
Knott, J. and G. Green (1976). "The initiation and propagation of ductile fracture in low strength
steels." Journal of Engineering Materials and Technology: 37.
Koplik, J. and A. Needleman (1988). "Void growth and coalescence in porous plastic solids."
International Journal of Solids and Structures 24(8): 835-853.
Kuna, M. and D. Sun (1996). "Three-dimensional cell model analyses of void growth in ductile
materials." International Journal of Fracture 81(3): 235-258.
Landron, C. (2011). INSA de Lyon.
Landron, C. (2011). Ductile damage characterization in Dual-Phase steels using X-ray
tomography, INSA de Lyon.
LauNielsen, K. and V. Tvergaard (2009). "Effect of a shear modified Gurson model on damage
development in a FSW tensile specimen." International Journal of Solids and Structures 46(3):
587-601.
Lee, S. and T. Su (1999). "Mechanical properties and microstructural features of AISI 4340 high-
strength alloy steel under quenched and tempered conditions." Journal of Materials Processing
Technology 87: 198-206.
Li, Y., et al. (2010). "Prediction of shear-induced fracture in sheet metal forming." Journal of
Materials Processing Technology 210(14): 1858-1869.
Page 175
160
Liu, B., et al. (2003). "The size effect on void growth in ductile materials." Journal of the
Mechanics and Physics of Solids 51(7): 1171-1187.
Liu, J., et al. (2013). "A stress-invariant based multi-parameters ductile progressive fracture
model." Materials Science and Engineering: A 576: 337-345.
Liu, Z., et al. (2016). "Void behaviors from low to high triaxialities: Transition from void collapse
to void coalescence." International Journal of Plasticity 84: 183-202.
Luo, M. and T. Wierzbicki (2010). "Numerical failure analysis of a stretch-bending test on dual-
phase steel sheets using a phenomenological fracture model." International Journal of Solids and
Structures 47(22): 3084-3102.
Maire, E., et al. (2008). "Initiation and growth of damage in a dual-phase steel observed by X-ray
microtomography." Acta Materialia 56(18): 4954-4964.
Malvern, L. E. (1969). Introduction to the Mechanics of a Continuous Medium.
McClintock, A. (1968). "A Criterion for Ductile Fracture by the Growth of Holes." Journal of
Applied Mechanics 35(2): 363.
McClintock, F. A. (1968). "A Criterion for Ductile Fracture by the Growth of Holes." Journal of
applied mechanics 35(2): 363.
Nahshon K, H., JW. (2008). Eur J Mech β A/Solids 27(1): 1β17.
Nahshon, K. and J. Hutchinson (2008). "Modification of the Gurson model for shear failure." Eur
J Mech β A/Solids 27(1): 1β17.
Nielsen, K. L. and V. Tvergaard (2009). "Effect of a shear modified Gurson model on damage
development in a FSW tensile specimen." International Journal of Solids and Structures 46(3):
587-601.
O'Brien, J. and G. D (1991). "Heat Treating, Plasma (Ion) Nitriding, MetalsHandbook." ASM 4.
Odeshi, G. and N. Bassim (2009). "High strain-rate fracture and failure of a high strength low alloy
steel in compression." Materials Science and Engineering: A 525(1-2): 96-101.
Pang, C., et al. (2013). "General relation between tensile strength and fatigue strength of metallic
materials." Materials Science and Engineering: A 564: 331-341.
Page 176
161
Pardoen, T. and J. Hutchinson (2000). "An extended model for void growth and coalescence."
Journal of the Mechanics and Physics of Solids 48(12): 2467-2512.
Podder, B., et al. (2012). "Effect of preform heat treatment on the flow formability and mechanical
properties of AISI4340 steel." Materials & Design 37: 174-181.
Puttick, K. (1959). "Ductile fracture in metals." Philosophical magazine 4(44): 964-969.
Qiao, Y., et al. (2017). "Study on coexistence of brittle and ductile fractures in nano reinforcement
composites under different loading conditions." International Journal of Fracture 204(2): 205-224.
Rice, J. and D. Tracey (1969). "On the ductile enlargement of voids in triaxial stress fields." Mech
Phys Solids 17(3): 201β217.
Rice JR, T. D. (1969). "On the ductile enlargement of voids in triaxial stress fields." Mech Phys
Solids 17(3): 201β217.
Rogers, H. (1960). "The tensile fracture of ductile metals." AIME trans 218(3): 498-506.
Shu, J. Y. (1998). "Scale-dependent deformation of porous single crystals." International Journal
of Plasticity 14(10): 1085-1107.
Simulia, D. S. (2017). "Abaqus/Standard Userβs Manual, Version 6.17 (2017)."
Sirin, S., et al. (2013). "Influence of initial conditions on the mechanical behavior of ion nitrided
AISI 4340 steel." Materials Science and Engineering: A 564: 232-241.
Slimane, A., et al. (2015). Journal of Materials Research and Technology 4(2): 217-223.
StΓ₯hlberg, U. (1986). "Influence of spread and stress on the closure of a central longitudinal hole
in the hot rolling of steel." Journal of mechanical working technology 13(1): 65-81.
Sun, D. Z., et al. (1989). "Application of local damage models to the numerical analysis of ductile
rupture." Fatigue & Fracture of Engineering Materials & Structures 12(3): 201-212.
Tanaka, K. and J. Spretnak (1973). "An analysis of plastic instability in pure shear in high strength
AISI 4340 steel." Metallurgical Transactions 4(2): 443.
Thomason, P. (1985). "A three-dimensional model for ductile fracture by the growth and
coalescence of microvoids." Acta Metallurgica 33(6): 1087-1095.
Page 177
162
Thomason, P. (1990). "Ductile fracture of metals." Pergamon Press plc, Ductile Fracture of
Metals(UK), 1990: 219.
Thompson, A. W. (1987). "Modeling of local strains in ductile fracture." Metallurgical and
Materials Transactions A 18(11): 1877-1886.
Tipper, C. (1949). "The fracture of metals." Metallurgia 39(231): 133-137.
Tresca, H. (1864). "Memoir on the flow of solid bodies under strong pressure." Comptes-rendus
de lβacadΓ©mie des sciences 59: 754-758.
Tvergaard, V. (1981). "Influence of voids on shear band instabilities under plane strain
conditions." International Journal of Fracture 17(4): 389-407.
Tvergaard, V. (1982). "Influence of void nucleation on ductile shear fracture at a free surface."
Journal of the Mechanics and Physics of Solids 30(6): 399-425.
Tvergaard, V. (1990). "Advances in Applied Mechanics." Academic Press, Inc 27: 83.
Tvergaard, V. and A. Needleman (1984). "Analysis of the cup-cone fracture in a round tensile
bar." Acta metallurgica 32(1): 157-169.
Wang, A., et al. (1996). "A study of pore closure and welding in hot rolling process." Journal of
Materials Processing Technology 60(1-4): 95-102.
Wang, H., et al. (2007). Crack and void healing in metals. Self Healing Materials, Springer: 255-
277.
Weck, A. G. (2007). The role of coalescence on ductile fracture.
Wei, D., et al. (2004). "Simulation of crack healing in BCC Fe." Scripta materialia 51(6): 583-587.
Wen, J., et al. (2005). "The modified Gurson model accounting for the void size effect."
International Journal of Plasticity 21(2): 381-395.
Wen, J., et al. (2005). "Extension of the gurson model accounting for the void size effect." Acta
Mechanica Sinica 21(2): 142-150.
Wierzbicki, T., et al. (2005). "Calibration and evaluation of seven fracture models." International
Journal of Mechanical Sciences 47(4): 719-743.
Page 178
163
Wilkins, M., et al. (1980). Cumulative-strain-damage model of ductile fracture: simulation and
prediction of engineering fracture tests, Lawrence Livermore National Lab., CA (USA); Science
Applications, Inc., San Leandro, CA (USA).
Wilsdorf, H. G. (1975). "Void initiation, growth, and coalescence in ductile fracture of metals."
Journal of Electronic Materials 4(5): 791-809.
Xue, L. (2008). "Constitutive modeling of void shearing effect in ductile fracture of porous
materials." Engineering Fracture Mechanics 75(11): 3343-3366.
Xue, L. and T. Wierzbicki (2008). "Ductile fracture initiation and propagation modeling using
damage plasticity theory." Engineering Fracture Mechanics 75(11): 3276-3293.
Xue, Z., et al. (2010). "Calibration procedures for a computational model of ductile fracture."
Engineering Fracture Mechanics 77(3): 492-509.
Zhang, K., et al. (1999). "Ductile fracture of materials with high void volumefraction."
International Journal of Solids and Structures 36(23): 3407-3425.
Zhang, S. and K. Hsia (2001). "Modeling the fracture of a sandwich structure due to cavitation in
a ductile adhesive layer." TRANSACTIONS-AMERICAN SOCIETY OF MECHANICAL
ENGINEERS JOURNAL OF APPLIED MECHANICS 68(1): 93-100.
Zhang, X.-X., et al. (2009). "A criterion for void closure in large ingots during hot forging." Journal
of Materials Processing Technology 209(4): 1950-1959.
Zhang, Z. (1994). A practical micro-mechanical model-based local approach methodology for the
analysis of ductile fracture of welded T-joints. Lappeenranta University of Technology, Finland,
Ph. D. Thesis.
Zhang, Z. and E. Niemi (1994). "A new failure criterion for the Gurson-Tvergaard dilational
constitutive model." International Journal of Fracture 70(4): 321-334.
Zhang, Z., et al. (2000). "A complete Gurson model approach for ductile fracture." Engineering
Fracture Mechanics 67(2): 155-168.
Zhao, J. (2000). "Applicability of MohrβCoulomb and HoekβBrown strength criteria to the
dynamic strength of brittle rock." International Journal of Rock Mechanics and Mining Sciences
37(7): 1115-1121.